AI in IA

To what extent and capacity can Artificial

Intelligence assist in International Arbitration
procedures and proceedings?

Terry Michael Rauch II, MBA, LLM

Contents
Abstract......................................................................................................................................................... 4
1. Introduction .......................................................................................................................................... 5
Background ............................................................................................................................................... 5
Study Aims ................................................................................................................................................ 6
2. Literature Review ................................................................................................................................. 7
International Arbitration in the Technology Age....................................................................................... 7
Infographic: International Arbitration Institutions and Bodies ......................................................... 9
Procedures, Frameworks, and Regulations of International Arbitration ............................................ 10
Concerns and Challenges Related to International Arbitration .......................................................... 12
The Intersection of AI and International Arbitration .............................................................................. 13
AI Techniques in International Arbitration .......................................................................................... 14
Infographic: Example Individual AI Roles in Arbitration Case Analysis ........................................... 18
Current Applications in International Arbitration ............................................................................... 18
AI Concerns ......................................................................................................................................... 21
Gaps in the Literature ............................................................................................................................. 22
3. Methodology ...................................................................................................................................... 23
Research Design ...................................................................................................................................... 23
Comparative Analysis .............................................................................................................................. 23
Data Collection Methods ........................................................................................................................ 24
Data Analysis ........................................................................................................................................... 24
Ethics, Trustworthiness and Credibility of the Research ......................................................................... 25
Limitations............................................................................................................................................... 25
4. Findings ............................................................................................................................................... 25
AI Advancements in International Arbitration ........................................................................................ 26
General Advantages of AI in International Arbitration ....................................................................... 32
Limitations of AI in International Arbitration ...................................................................................... 33
Future AI and International Arbitration .............................................................................................. 36
Ethical and Legal Challenges Associated with AI in International Arbitration ........................................ 38
Infographic: AI Values and Principles .............................................................................................. 38
AI Bias and the Implications for Fairness ............................................................................................ 38
Infographic: Various forms of AI bias .............................................................................................. 40
AI's Limitations in Recognizing Cultural Differences ........................................................................... 41
 Accountability and Transparency Challenges in AI Decisions ............................................................. 43
Data Privacy and Security Concerns in AI Systems.............................................................................. 44
Regulatory Challenges in AI Deployment ............................................................................................ 46
Infographic: US State Legislations on AI ......................................................................................... 48
Summary of Findings............................................................................................................................... 50
5. Recommendations.............................................................................................................................. 51
Limiting AI Bias ........................................................................................................................................ 51
Improving Accountability and Transparency ........................................................................................... 52
Ensuring Data Privacy.............................................................................................................................. 53
Enhancing AI Governance and Regulation .............................................................................................. 54
6. Conclusion .......................................................................................................................................... 55
7. References .......................................................................................................................................... 58
Cases ....................................................................................................................................................... 58
Statutes, Legislation, Cases, and Legal Instruments ............................................................................... 58
Books and Reports .................................................................................................................................. 59
Journal Articles ........................................................................................................................................ 60
Websites and Other Online Sources ....................................................................................................... 62
Abstract
The growing integration of Artificial Intelligence (AI) across the legal sector raises questions
about its applicability and effectiveness, particularly in cross-border international arbitration. This
dissertation delves into the extent and capacity to which AI is currently and, in the future, can further
assist across the broader range of procedural and processual elements of international arbitration.
Beginning with a comprehensive literature review, the study investigates the foundation of international
arbitration and its associated processes, challenges, and the intersection with AI, including current tools,
applications, and concerns. Utilizing a broad legal analysis, the research examines how advancements in
AI along with global events drive new processes and procedures in International Arbitration.

Findings reveal several advantages of AI in international arbitration, such as increased efficiency,

accuracy, consistency, cost-effectiveness, and clarity. However, challenges emerge in AI's comprehension
of complex legal notions and nuanced human emotions. Significant concerns include decision
interpretability, data security, and ethical challenges like AI bias, transparency, and accountability in
decision-making. Furthermore, the study highlights the potential roles of large language models,
emotion AI, and the potential implications of Artificial General Intelligence (AGI) in international
arbitration.

Based on the findings, this research offers legal recommendations to help mitigate AI bias,
enhance accountability and transparency, ensure robust data privacy, and improve AI regulations in the
context of international arbitration. The study concludes by emphasizing a balanced approach,
integrating AI's capabilities with human expertise, to harness its full potential in international arbitration
while addressing inherent challenges.
1. Introduction
Background
Alternative Dispute Resolution (ADR) methods, such as mediation and arbitration, have become
increasingly popular as alternatives to traditional litigation. In particular, international arbitration plays a
significant and growing role in reducing court caseloads and helps courts to operate more efficiently,
focusing on the most complex and contentious issues that may not be suitable for ADR mechanisms.1

While ADR methods have been instrumental in streamlining the legal process, there's another
transformative force reshaping the legal landscape: the rise of Artificial Intelligence (AI). The use and
diversity of AI in the legal sector is growing, ranging from predicting outcomes to automating legal
research and contract analysis. The incorporation of AI into the legal sector is a logical continuation of
the use of science and technology in law, following trends such as the use of forensic science in the
courtroom, the creation of databases for legal statutes and case law, and the development of online legal
services. Just as these technological advancements transformed the practice of law, AI is poised to do the
same.2

The legal industry, traditionally slower in adopting innovative technologies due to regulatory
constraints and the intrinsic complexity of law, has begun to open its gates to AI and machine learning.
AI’s reach has extended into international dispute resolution, a field heavily reliant on human judgment
and expertise. This paper explores the current and future applications of AI in international arbitration
and analyzes the implications on practice and policy.

AI’s ability to process large volumes of information, recognize patterns, and predict outcomes
can make it an invaluable tool for resolving disputes. Traditionally, courts have been strained with
overburdening caseloads that cause delays and inefficiencies in the resolution of disputes.

In recent years, AI has begun to make inroads in the realm of dispute resolution. For instance,
EviSort AI Labs has created an AI-powered research and contract analysis tool that has been adopted by
numerous law firms and legal departments. Evisort offers a streamlined contract management solution
powered by artificial intelligence.3 Similarly, Premonition, an AI litigation analytics tool, uses Big Data and

1
T J Stipanowich, 'Arbitration: The "New Litigation"' (2010) 1 University of Illinois Law Review 1.
2
P Leith and A Hoey, The Computerised Lawyer: a guide to the use of computers in the legal profession (Springer
Science & Business Media 2012).
3
N Yamane, 'Artificial Intelligence in the Legal Field and the Indispensable Human Element Legal Ethics Demands'
(2020) 33 Geo J Legal Ethics 877.
machine learning to predict case outcomes based on historical data, including the judge’s past decisions
and the win/loss record of attorneys. These AI applications illustrate the potential of AI in making legal
processes more efficient and insightful.

AI is also offering new pathways for dispute resolution. Online Dispute Resolution (ODR)
platforms incorporate AI to facilitate negotiation, mediation, and arbitration in an online environment,
thereby making dispute resolution more accessible and less time-consuming.4 Furthermore, with
advancements in Natural Language Processing and argumentation mining, AI systems are being
developed to simulate human-like negotiation, potentially offering novel means of resolving disputes.5

Despite these advancements, there is a need to closely examine the ethical, legal, and practical
implications of AI, particularly in International or cross border disputes. Questions related to
transparency, accountability, and fairness of AI systems, as well as the risk of AI-induced job
displacement in the legal profession, need careful deliberation. Moreover, concerns over AI system
reliability and security remain contentious issues.

Study Aims
The objective of this paper is to investigate the current applications of AI in international
disputes, its potential in alleviating court burden, and the accompanying challenges. Through a
comprehensive analysis of arbitral awards, case law, and empirical research, this paper will shed light on
how AI might shape the future of international arbitration, specifically in the context of arbitration.

Main question: To what extent and capacity can Artificial Intelligence assist in international
arbitration proceedings?

In addressing the main question, this dissertation is not merely focused on AI's ability to assist as
an arbitrator but is rather taking a holistic view of the arbitration landscape, encompassing a broader
range of procedural and processual elements. The objective is to paint a comprehensive picture of the
myriad ways AI can augment or otherwise support the intricate mechanisms of international arbitration.
Therefore, several additional questions will be addressed including:

1. What are the key features of AI that are, and can in the future, contribute to the resolution of
legal disputes?

4
M Moscati and others, Comparative Dispute Resolution (Edward Elgar 2020) 339.
5
S Dinnar and others, "Artificial intelligence and technology in teaching negotiation" (2021) 37(1) Negotiation
Journal 65.
 2. What are current and inherent limitations of AI that could interfere with fair and effective
arbitration procedures?
3. What are the potential ethical and legal challenges associated with AI use in international
arbitration?
4. Are there successful real-world examples of AI implementation in international arbitration?

2. Literature Review
Arbitral systems around the world have witnessed an unprecedented upsurge in the volume of
cases, contributing to increased court burdens. A growing percentage of these are cross border disputes
due to the growth of international trade and transactions.

Contemporary literature points towards the potential of Artificial Intelligence (AI) in alleviating
this issue. However, to truly grasp the potential impact and opportunities presented by AI, it is important
to first explore the intricacies of its Processes and Procedures, Legal Framework, and Regulations of
International Arbitration. These foundations dictate not only how parties approach and navigate
disputes, but also how arbitral tribunals ensure fair and equitable outcomes. As these processes and
legal standards evolve, so too do the opportunities for AI integration. Reviewing these changes and
advancements is pivotal, as they provide a lens through which one can forecast the trajectory of
International Arbitration. This understanding subsequently enables us to ascertain how AI can be
harnessed to optimize efficiency and bolster the efficacy of outcomes in International Arbitration.

International Arbitration in the Technology Age

Between 1980 and 2020, international arbitration witnessed a remarkable expansion. Major
arbitration institutions, including the ICC, LCIA, and SIAC, experienced a significant rise in administered
cases. This surge can be attributed to the global evolution of trade and the intricate nature of
international business dealings.6

With the inception of the digital age, technological innovations have deeply influenced the
manner, efficiency, and accessibility of international business. In turn, the late 20th and early 21st
centuries also saw significant developments in the law governing international arbitration. The

6
ICC, 'ICC Dispute Resolution Statistics' <https://iccwbo.org/publication/icc-dispute-resolution-statistics/>; LCIA,
'LCIA Facts and Figures - 2019' <https://www.lcia.org/News/lcia-facts-and-figures-2019.aspx>; SIAC, 'SIAC Annual
Report 2019' <https://www.siac.org.sg/2015-11-09-01-49-41/facts-figures/annual-report all> accessed 30 July
2023.
UNCITRAL Model Law on International Commercial Arbitration, adopted in 1985, has been influential in
shaping the arbitration laws of many countries.7 Also known as the New York Convention, this
convention has been ratified by more than 160 countries as of 2020, facilitating the enforcement of
arbitral awards across borders.8

While most arbitration cases involve commercial contract disputes, there are three primary
types of international arbitration, with each type catering to specific participants and the nature of their
disagreements.

• Interstate Arbitration: This involves nations represented by their governments, resolving

disputes through arbitration. While it is fundamentally a legal process, its significance extends
beyond just legal terms.
• Investor-State Arbitration: This type of arbitration involves disputes between the nations and
the private sector, such as foreign nationals or companies. Originating from bilateral and
multilateral investment treaties, this form of arbitration allows foreign investors to seek
compensation from host governments.
• International Commercial Arbitration: This is the most prevalent form of international
arbitration and involves parties from different countries, typically concerning contractual
disputes between corporations. Businesses often prefer this method over litigation in national
courts, believing that an international tribunal is more impartial and better informed about
international business practices. Most contracts between corporations from different nations
include a clause specifying that disputes will be resolved through arbitration.

Scholars have extensively explored the role and importance of these international arbitration
mechanisms, creating a rich academic debate. In particular, the literature shows a substantial focus on
issues related to the partiality of national courts in handling cross-border disputes9 and, in turn, the
significance of international arbitration as a neutral medium (an issue discussed later in this paper in
relation to AI).

7
UNCITRAL, 'Model Law on International Commercial Arbitration 1985'
<https://uncitral.un.org/en/texts/arbitration/modellaw/commercial_arbitration> accessed 30 July 2023.
8
UNCITRAL, 'Status: Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York, 1958)'
<https://uncitral.un.org/en/texts/arbitration/conventions/foreign_arbitral_awards/status2> accessed 30 July 2023.
9
J Paulsson, 'Arbitration in Three Dimensions' (2010) 60 International and Comparative Law Quarterly 291.
 Some scholars have highlighted the inherent procedural flexibility of international arbitration,
positing that it serves as a more efficient alternative to conventional court litigation.10 While others
contend that, while arbitration can offer efficiency, it doesn't guarantee it, noting cases where arbitration
becomes costly and protracted.11 Efficiency and cast are additional issues discussed later in this paper in
relation to the use of AI in arbitration.

The issue of enforceability in international arbitration is another theme recurrent in literature.

Scholars emphasize the pivotal role of the New York Convention in furnishing a sturdy global structure
for arbitral award enforcement.12 In the realm of investor-state disputes, works like those of Dolzer and
Schreuer illuminate the importance and wide availability of arbitration institutions, particularly due to its
reputed neutrality and expertise in international investment law.13 The InfoGraphic below show the
primary institutions and bodies in International Arbitration.

Infographic: International Arbitration Institutions and Bodies

10
G Born, International Commercial Arbitration (2nd edn, Kluwer Law International 2014).
11
S Brekoulakis, 'On Arbitrability: Persisting Misconceptions and New Areas of Concern' in L Mistelis & S Brekoulakis
(eds), Arbitrability: International and Comparative Perspectives (Kluwer Law International 2009).
12
J Lew, L Mistelis and S Kröll, Comparative International Commercial Arbitration (Kluwer Law International 2003).
13
R Dolzer and C Schreuer, Principles of International Investment Law (2nd edn, Oxford University Press 2012).
Procedures, Frameworks, and Regulations of International Arbitration
Across these institutions, the decision-making process hinges upon ensuring fairness,
impartiality, and enforceability. Arbiters within these institutions must be flexible when balancing
honoring party autonomy while also adhering to justice and due process. To guarantee just and
equitable outcomes, there are specific guidelines and procedures arbitrators are expected to follow.14
Following these guidelines is critical to ensure that awards are final, binding, and enforceable.

While the foundational procedures and guidelines remain essential for the legitimacy of
decisions, the structural elements of international arbitration are equally significant.

International arbitration is grounded in four principal components; the arbitration agreement,

the arbitral tribunal, the seat of arbitration, and the arbitration decision.15 These components operate
within, and acquire their validity and enforceability, from legal frameworks and regulations which
frequently undergo modifications considering newly established conventions, treaties, and national
developments. The primary objectives behind these updates are to enhance efficiency, transparency, and
accessibility —objectives that align well with the benefits provided by the integration of AI.

An example of recent changes in the international arbitration framework can be seen in the 2019
Singapore Convention on Mediation in 2019.16 This Convention, ratified by multiple countries, has
transformed how cross-border commercial disputes are settled. It defines 'mediation' in Art. 2(3) as
including both facilitative and advisory methods but excludes arbitration and adjudication. Notably, it
recognizes mediation techniques powered by artificial intelligence algorithms.17

Additionally, it should be noted that all UNESCO member states agreed on a global standard for
AI ethics.18 This standard ensures AI's development and use follows the rule of law, avoids harm, and
offers redress for the affected. It emphasizes transparency, accountability, and privacy across sectors

14
G Born, International Commercial Arbitration (2nd edn, Kluwer Law International 2014).
15
M Moscati and others, Comparative Dispute Resolution (Edward Elgar 2020) 8.
16
United Nations Convention on International Settlement Agreements Resulting from Mediation (signed 7 August
2019, entered into force [date of entry into force]) UN Doc V1900316
<https://uncitral.un.org/sites/uncitral.un.org/files/media-
documents/EN/Texts/UNCITRAL/Arbitration/mediation_convention_v1900316_eng.pdf>
17
Ibid art 2(3).
18
UNESCO, 'Recommendation on the Ethics of Artificial Intelligence' (2022)
<https://unesdoc.unesco.org/ark:/48223/pf0000381137>.
such as law, education, healthcare, and economics. Furthermore, it promotes stronger data protection,
granting people and organizations more control over their data.19

Similarly, changes in national arbitration laws have also played a pivotal role in shaping the
international arbitration landscape. For instance, the United Kingdom has recently amended its
Arbitration Act 1996 to enhance the efficiency and cost-effectiveness of arbitration proceedings.20 The
amendments include provisions for emergency arbitrator procedures and expedited arbitration
proceedings.

The United States has updated its Federal Arbitration Act to provide greater clarity on the scope
of judicial review of arbitration awards.21 The amendments aim to strike a balance between the need for
finality of arbitration awards and the need to ensure that the awards are not tainted by serious
procedural irregularities.

Singapore has revised its International Arbitration Act to include provisions for third-party
funding of arbitration.22 This is a significant development, as it allows parties with limited resources to
pursue their claims through arbitration.

Many arbitration institutions also quickly adapted their rules and procedures in response to the
COVID-19 pandemic.23 For instance, The International Chamber of Commerce (ICC) issued guidance on
conducting virtual hearings and made changes to its rules to facilitate the use of electronic documents.24

As the arbitration institute and bodies continually refine the legal frameworks and regulations
governing international arbitration, the goal remains to bolster the system's efficiency, transparency, and
accessibility. However, to understand AI ability to assist these institutions, it is also crucial to understand
current concerns and challenges related to international arbitration.

19
Ibid.
20
UK Government, 'Arbitration Act 1996 (Amendment)' (2022)
<https://www.legislation.gov.uk/ukpga/2022/15/contents/enacted> accessed 30 July 2023.
21
US Government, 'Federal Arbitration Act (Amendment)' (2022) <https://www.congress.gov/bill/117th-
congress/house-bill/963/text> accessed 3 July 2023.
22
Singapore Government, 'International Arbitration (Amendment) Act' (2021) <https://sso.agc.gov.sg/SL/S971-
2020?Timeline=On> accessed 3 July 2023.
23
T Landon and K von der Weid, 'The Impact of COVID-19 on International Arbitration Procedure' in T Landon and K
von der Weid (eds), The Impact of Covid on International Disputes (2022) 84.
24
ICC, 'Leveraging Technology for Fair, Effective and Efficient International Arbitration Proceedings' (ICC Publication,
18 February 2022).
Concerns and Challenges Related to International Arbitration
While the advantages of privacy and confidentiality in arbitration are well-documented,25 current
trends show arbitrations are increasingly deciding matters of notable public importance. Therein lies a
dilemma: the privacy that attracts parties to arbitration might work against the public good in certain
instances.

Numerous scholars have advanced the idea that greater transparency will fortify the integrity of
the arbitration and foster consistency in decisions by considering prior awards as potential precedents.26
The public's trust in the mechanism might also be enhanced if they are granted a clearer view into its
workings.

The growing importance of transparency which resulted in international instruments like the
UNCITRAL Transparency Rules27 and the Mauritius Convention on Transparency28 to be established.
Additionally, arbitration bodies like the ICC and LCIA, as well as national legislations29, have made
amendments promoting transparency, especially when public interests are implicated.

This is important when considering AI’s role in international arbitration as access to case and
award data is required for developing and training AI algorithms to aid in arbitration.30 However, while
transparency is one facet, other concerns exist in the arbitration landscape - these include questions
over the legitimacy of arbitral tribunals and the enforcement of awards across jurisdictions. The debates
among scholars and practitioners revolve around whether such challenges are inherent to and
surmountable within the process, or if they signify a need for systemic reforms in international
arbitration.

Integrating AI tools in this field raises distinct questions, and it is crucial that technology
augments rather than impedes the arbitration process, addressing existing concerns rather than

25
M Ristovska, 'The Principle of Confidentiality In International Arbitration' (2021) 15(2) IBANESS Conference
Series-Plovdiv/Bulgaria 415-421.
26
L Nottage, 'Confidentiality and Transparency in International Arbitration: Asia-Pacific Tensions and Expectations'
(2020) 16(1) Asian International Arbitration Journal.
27
UNCITRAL, 'Transparency in Treaty-based Investor-State Arbitration' (2013)
<https://uncitral.un.org/en/texts/arbitration/contractualtexts/transparency.
28
United Nations, 'United Nations Convention on Transparency in Treaty-based Investor-State Arbitration' (2015)
<https://uncitral.un.org/en/texts/arbitration/conventions/transparency>.
29
'Arbitration Laws of the World' (International Arbitration Information, 8 July 2023) <https://www.international-
arbitration-attorney.com/arbitration-law-of-world/> accessed 25 September 2023.
30
F Spyropoulos and E Androulaki, 'Aspects of Artificial Intelligence on E-Justice and Personal Data Limitations'
(2023) 26(3) Journal of Legal, Ethical and Regulatory Issues 1-8.
exacerbating them. The literature emphasizes the importance of striking an appropriate balance around
confidentiality, which is seen as both a shield enabling free expression and a veil leading to
unpredictability due to inaccessible precedent.

The perceived neutrality of international arbitration, facilitated by the ability to select arbitrators
and applicable law, is also often challenged. Concerns are raised about possible biases stemming from
arbitrators' affiliations and the lack of diversity among arbiter pools.31

Procedural fairness and flexibility in arbitration are double-edged, with scholars arguing that
while it can enhance efficiency, the absence of strict rules can compromise fairness and result in differing
standards. Likewise, the importance of cross-cultural understanding in international arbitration enables a
diverse range of perspectives but also introduces risks of moral equivalence which could disregard the
hierarchical weight of norms and moral imperatives.32 Lastly, the accessibility of international arbitration
remains debated, with high costs and complex processes potentially barring some parties and
developing nations from opting for this dispute resolution method.33

Having reviewed the literature on the intricacies, challenges, and evolution in international
arbitration, we are provided a context with which to explore the literature on the intersection of AI and
International Arbitration.

The Intersection of AI and International Arbitration

The term "AI" was first coined by John McCarthy in 1956 at the Dartmouth Conference, where
he defined it as "the science and engineering of making intelligent machines".34 This definition, however,
is broad and has been subject to various interpretations.

31
Tumey v Ohio 273 U.S. 510 (1927) (noting that "Arbitrators are not national court judges; they are usually private
practitioners, of some sort, engaged in the business of providing legal services for a fee. Often, they face significant
financial and competitive pressure to earn more money and handle more cases").
32
L Rosen, 'Law as Culture: An Invitation' (2011) 3 McGeorge School of Law Global Center for Business and
Development Annual Symposium.
33
F Fortese and L Hemmi, 'Procedural Fairness and Efficiency in International Arbitration' (2015) 3(1) Groningen
Journal of International Law.
34
J McCarthy, M Minsky, N Rochester and C Shannon, 'A Proposal for the Dartmouth Summer Research Project on
Artificial Intelligence' AI Magazine, 27(4), 12 (1955).
 In contrast, the Stanford Encyclopedia of Philosophy defines AI as "the use of computers to
simulate human intelligence".35 This definition emphasizes the human-like capabilities of AI, including
learning, reasoning, problem-solving, perception, and language understanding.

AI Techniques in International Arbitration

To fully understand the growing influence of AI tools in International Arbitration it is critical to
first review the overarching roles of Information Technology (IT) and digitalization in international
arbitration. IT revolutionized the ways in which arbitration proceedings are conducted, including the way
evidence is collected, stored, and shared amongst the parties. Through digital platforms, arbitrators and
parties exchange documents and information, collaborate remotely and conduct online hearings.36

IT and digital data have had significant impacts on the efficiency, accessibility, and transparency
of arbitration. Electronic document management systems have resulted in decreased paperwork,
enhanced accessibility, and increased efficiency in managing the arbitration process.37 Furthermore,
online case management platforms have enabled parties to track the progress of arbitration proceedings
in real time, thereby promoting transparency.

These advancements in IT serve as the foundational infrastructure for AI, providing the
hardware, software, and data frameworks essential for developing and deploying AI algorithms and
solutions.

Robotic Process Automation: Robotic Process Automation (RPA) is a form of AI technology

enabling software bots to perform high-volume, repetitive tasks by mimicking human actions. It’s widely
applied in international arbitration to automate tasks like legal billing, contract review, and discovery,
yielding significant time and cost savings. RPA is beneficial in handling routine tasks like data entry and
document review across various systems within organizations, enhancing efficiency, accuracy, and
compliance by reducing human intervention.38 It aids in structuring and managing contract data
meticulously, minimizing the risk of errors and legal disputes. Although primarily used for rule-based

35
B Copeland, 'Artificial Intelligence' in E N Zalta (ed), The Stanford Encyclopedia of Philosophy (Winter 2019
Edition) (2019).
36
M S Karim, 'Artificial Intelligence: An Undiscovered Future of Arbitration' (2019) 22(2) Int A.L.R. 47.
37
P Leith and A Hoey, The Computerised Lawyer: a guide to the use of computers in the legal profession (Springer
Science & Business Media 2012).
38
UiPath Newsroom, 'UiPath Automates Facility for Conclusion of Arbitration Agreements to Support Business
Continuity for Companies, Arbitration Courts' (9 June 2020) <https://www.uipath.com/newsroom/uipath-
automates-facility-for-conclusion-of-arbitration-agreements-to-support-business-continuity-amid-crisis-for-
companies-arbitration-courts> accessed 2 August 2023.
tasks, RPA can integrate with advanced AI tools to develop systems capable of learning, decision-making,
and understanding human language, allowing for greater complexity and application in various business
processes.

RPA Technology Example: UiPath, a leading Robotic Process Automation (RPA) enterprise
software company, collaborated with the Bucharest International Arbitration Court (BIAC) to launch the
world's first fully automated institutional mediation process for concluding arbitration agreements. The
system extracts relevant data from parties, generates the Arbitration Agreement, and gathers consent
via DocuSign, ensuring a secure and reliable electronic process.39

Machine Learning: Machine Learning (ML) often builds upon more basic forms of AI such as RPA.
It is used in various aspects of International Arbitration. One of the most prominent is predictive
modeling and coding, which aids in spotting pertinent documents by training algorithms with human-
labeled data, categorizing them as relevant or not. Using ML for document review and discovery - areas
notoriously time-consuming and resource-intensive. - delivers time and cost efficiencies and increased
accuracy in arbitration.40

Predictive modeling is being applied to several large databases containing historical data to
discern outcome patterns, providing insights for similar future disputes. The repeated ability to
accurately forecast arbitration decisions using machine learning is nascent but promising. With inputs
like dispute nature, jurisdiction, involved arbitrators, and legal issues, algorithms are currently able to
hint at probable results.41

ML Technology Example: Tools such as Veritone's aiWARE use continuous active learning to code
and review large legal data sources. While coding and providing outputs, aiWARE also self-improves
enhancing its own review accuracy without needing any further input.42 Similarly, eBrevia’s

39
Ibid.
40
C M de Westgaver, 'Canvassing views on AI in IA: The Rise of Machine Learning' (Kluwer Arbitration Blog, 12 July
2023) <https://arbitrationblog.kluwerarbitration.com/2023/07/12/canvassing-views-on-ai-in-ia-the-rise-of-
machine-learning/> accessed 17 August 2023.
41
A Wilkinson, J Tsai and D Curle, 'How natural language processing can improve legal search results' (Kira Systems,
202AD) <https://kirasystems.com/learn/how-natural-language-processing-improving-can-improve-legal-search-
results/#:~:text=NLP%20and%20machine%20learning%20can,rule%20on%20a%20given%20case> accessed 6
August 2023.
42
D Wong, 'Technology Assisted Review: How AI and Machine Learning Streamline the Review Process' (Veritone, 5
May 2023) <https://www.veritone.com/blog/technology-assisted-review/> accessed 6 September 2023.
ContractTracker uses RPL and ML to automate the contract review process, thereby improving the
accuracy and speed of document review.43

Natural Language Processing (NLP): Natural Language Processing (NLP) is a field of artificial
intelligence that focuses on the interaction between computers and humans through natural language. It
has demonstrated value in various areas of legal practice, including Legal Research, Legal Document
Analysis, Predictive Modelling, Machine Translation, and Sentiment Analysis.44 NLP enhances search
capabilities by comprehending the underlying meaning, synonyms, antonyms, and even the context of
the entered search query.45 It can identify specific data from texts, such as names, dates, obligations, and
more. This is especially valuable in legal scenarios where specific clauses, terms, patterns, and trends can
be identified.

NLP is also used to overcome language barriers. It can automatically translate legal documents
into various languages, thus facilitating the global interaction and understandings during arbitration.46
NLP is also used in content sentiment analysis to help understand the emotional undertones of legal
texts which could reveal biases and tendencies in decision-making.47

NLP Technology Example: Lex Machina, a legal analytics platform, uses NLP to process and
analyze legal documents, assisting attorneys in forecasting outcomes based on historical data.48
Similarly, platforms like Kira, ThoughtRiver, and Luminance use NLP to automatically review and extract
information from contracts and other legal documents, potentially highlighting clauses or patterns that
have previously led to arbitration.49

43
‘ContractTracker’ (Donnelley Financial Solutions
(DFIN)).<https://www.dfinsolutions.com/products/ebrevia/contracttracker> accessed 6 September 2023.
44
K Ashley, Artificial Intelligence and Legal Analytics: New Tools for Law Practice in the Digital Age (Cambridge
University Press 2017).
45
A Wilkinson, J Tsai and D Curle, 'How natural language processing can improve legal search results' (Kira Systems,
202AD) <https://kirasystems.com/learn/how-natural-language-processing-improving-can-improve-legal-search-
results/#:~:text=NLP%20and%20machine%20learning%20can,rule%20on%20a%20given%20case> accessed 6
August 2023.
46
R Raghunathan, 'How do we break the language barrier in NLP' (LatentView Analytics, 2020)
<https://www.latentview.com/blog/how-do-we-break-the-language-barrier-in-nlp/> accessed 12 August 2023.
47
Altlaw, 'What is sentiment analysis? using NLP in Ediscovery' (Altlaw eDiscovery, 2023)
<https://www.altlaw.co.uk/blog/what-is-sentiment-analysis-using-nlp-in-ediscovery> accessed 26 July 2023.
48
Lex Machina, 'Legal Analytics' <https://lexmachina.com> accessed 6 August 2023.
49
Litera, <https://www.litera.com>; ThoughtRiver, <https://www.thoughtriver.com>; Luminance,
<https://www.luminance.com> all accessed 6 August 2023.
 Generative AI, Neural nets, and Fuzzy Logic: Generative AI (GenAI) tools leverage deep learning,
to generate multilanguage human-like responses with precision, accuracy, and efficiency. With every
request these models learn by analyzing enormous volumes of text data, adjusting their understanding,
and enhancing their predictions and responses. GenAI is currently used to provide objective analysis,
reducing costs. These tools are particularly useful in identifying patterns and correlations in large data
sets, such as past arbitral award databases, to provide insights for current dispute resolution far beyond
human capabilities.50

Closely associated with GenAI, Neural net models emulate the human brain's neural networks to
learn and make predictions. These models can significantly impact dispute resolution by providing
insights on complex multifactorial cases.51

Fuzzy logic models incorporate elements of uncertainty in decision-making, mirroring the legal
system's often "grey" nature. This can provide a more nuanced understanding of arbitration cases,
allowing for the consideration of less definable factors.52

GenAI Technology Example: Tools such as ChatGPT can potentially serve as a tribunal secretary,
handling tasks like drafting procedural orders and summarizing proceedings. This could significantly
reduce the time and costs associated with non-decisional parts of an award. Of note, ChatGPT is already
finding its way in arbitration hearings. In 2023, a Colombian judge utilized ChatGPT in a health insurance
dispute, asking both decisional and research questions. Similarly, a Pakistani court and a company have
also explored ChatGPT's potential in legal scenarios.53

50
H Prakken, A Wyner, T Bench-Capon and K Atkinson, 'A formalization of argumentation schemes for legal case-
based reasoning in ASPIC+' (2015) 25 Journal of Logic and Computation 1141–1166.
51
I Chalkidis, I Androutsopoulos and N Aletras, 'Neural Legal Judgment Prediction in English' in Proceedings of the
57th Annual Meeting of the Association for Computational Li O Colorado, 'The Future of International Arbitration in
the Age of Artificial Intelligence' (2023) 40(3) Journal of International Arbitration 328.istics (Association for
Computational Linguistics, Florence 2019) 4317, <https://www.aclweb.org/anthology/P19-1424> accessed 17
August 2023.
52
N Bagherian-Marandi and M Ravanshadnia, 'Two-layered fuzzy logic-based model for predicting court decisions
in construction contract disputes' (2021) 29 Artificial Intelligence and Law 453.
53
S Efstathiou and M Apostol, 'Arbitration tech toolbox: Chatgpt - Arbitral assistant or fourth arbitrator?' (2023)
Kluwer Arbitration Blog <https://arbitrationblog.kluwerarbitration.com/2023/07/22/arbitration-tech-toolbox-
chatgpt-arbitral-assistant-or-fourth-arbitrator/> accessed 07 August 2023.
 Infographic: Example Individual AI Roles in Arbitration Case Analysis

Current Applications in International Arbitration

There have been several publications examining the specific applications of AI in arbitration
processes and procedures, evaluating both the opportunities and challenges. The primary uses are
discussed below.

Legal Research and Due Diligence: Legal research has been augmented by the advent of AI
technologies. AI can examine vast amounts of information far more efficiently than humans and can
deliver a set of results based on the parameters specified.54 A popular tool used in legal research is ROSS
Intelligence, an AI program that can answer legal questions in plain English. ROSS uses machine learning
to understand the complexities of law and to provide comprehensive research.55

AI has also been applied in due diligence tasks to decrease the substantial time investment that
traditional methods require. For example, Kira Systems is a machine learning software that assists with
due diligence by accurately identifying, extracting, and analyzing information from contracts and other
documents.56 AI-driven due diligence has been utilized in high-profile cases, such as JP Morgan's
deployment of COIN, a contract analysis tool that helped complete 360,000 hours of legal work in a
matter of seconds.57

Contract Review and Management: AI has made significant strides in contract review and
management. AI applications can process and understand contracts, summarizing them into digestible
information. This ability is particularly useful when dealing with large volumes of contracts or when in
need of quick reviews. For instance, LawGeex, a contract review platform, uses AI to ensure legal
compliance by analyzing contracts against predefined policies.58

AI has proven vital in contract management, offering insights that aid in making strategic
decisions. Tools such as eBrevia employ AI algorithms to extract and structure data from contracts for
easy review and interpretation.59 This ability is integral in maintaining contract compliance, risk
management, and efficient workflow.

Prediction of Legal Outcomes: AI has the potential to revolutionize legal outcome predictions.
Programs like Lex Machina use natural language processing and machine learning to analyze the
historical data of court decisions to predict future outcomes.60 Lex Machina and several other platforms
utilize machine learning algorithms to provide predictive analytics, offering insights into probable case

54
N Yamane, 'Artificial Intelligence in the Legal Field and the Indispensable Human Element Legal Ethics Demands'
(2020) 33 Geo J Legal Ethics 877.
55
N Yamane, 'Artificial Intelligence in the Legal Field and the Indispensable Human Element Legal Ethics Demands'
(2020) 33 Geo J Legal Ethics 877.
56
Kira Systems, https://kirasystems.com/ accessed 11 September 2023.
57
H Son, 'This Software Does in Seconds What Took Lawyers 360,000 Hours' (The Independent, 28 February 2017)
<www.independent.co.uk/news/business/news/jp-morgan-software-lawyers-coin-contract-intelligence-parsing-
financial-deals-seconds-legal-working-a7603256.html> accessed 14 August 2023.
58
‘Contract Review Automation’ (Lawgeex, 3 November 2021) https://www.lawgeex.com/cra/ accessed 9
September 2023
59
'Ebravia app' https://www.dfinsolutions.com/products/ebrevia accessed 15 August 2023.
60
Lex Machina, 'Legal Analytics' <https://lexmachina.com> accessed 6 August 2023.
resolutions, which can help in pre-trial negotiations and potentially decrease litigation volume.61 The
predictive power of AI has been demonstrated in the Supreme Court of the United States, where an AI
model achieved a 70.2% accuracy rate in predicting decisions from 1816 to 2015, which was only slightly
lower than expert predictions.62 This suggests that AI can serve as a powerful tool for legal practitioners
to strategize and make informed decisions.

Case Strategy and Arbiter Analytics: AI technology has transformed case strategy and legal
analytics by providing insights into arbiter and court trends. For instance, Arbitrator Intelligence is an AI-
powered platform that provides feedback from lawyers, details about procedural rulings, and
perspectives from arbitrators. The platform addresses the problem of confidential information in
arbitrator research by leveraging AI analytics and a global network of experts.63

Other tools such as Kluwer Arbitration's Practical Tools Platform analyze massive amounts of
legal data to uncover trends and patterns that can aid in case strategy.64 These applications have
increased the efficiency of arbitration practices and led to more effective legal strategies.

Online Dispute Resolution (ODR): Online Dispute Resolution (ODR) is another area in law where
AI applications are making a mark. AI facilitates ODR by automating the resolution process. This
application is exemplified by Modria, a platform which uses AI to help parties in e-commerce disputes
resolve their issues without needing a court or arbitrator.65

Specific evidence of AI's impact on ODR comes from British Columbia's Civil Resolution Tribunal,
the world's first online tribunal, which uses AI to assist in resolving small claims disputes and strata
property issues.66 The European Union (EU) has also promoted ODR, broadening the concept of access to
justice to include both judicial and extrajudicial methods. In 2016, the European Commission introduced
an online dispute resolution platform.67 Besides regulations, European courts also have

61
H Surden, 'Machine Learning and Law' (2014) 89 Washington Law Review 87.
62
D Katz, M Bommarito and J Blackman, 'A general approach for predicting the behavior of the Supreme Court of
the United States' PloS one, 12(4), e0174698.
63
‘Arbitrator Intelligence Platform’ (Arbitrator Intelligence | Home) https://arbitratorintelligence.vercel.app/
accessed 9 September 2023
64
‘Practical Tools Platform’ (Kluwer Arbitration)
<https://www.wolterskluwer.com/en/solutions/kluwerarbitration/practical-tools> accessed 11 September 2023
65
‘Online Dispute Resolution Powered by Modria’ (Tylertech) <https://www.tylertech.com/products/online-
dispute-resolution> accessed 11 September 2023
66
‘Civil Resolution Tribunal’ (British Columbia’s Civil Resolution Tribunal) <https://civilresolutionbc.ca/> accessed 11
September 2023
67
European Commission, 'Alternative and Online Dispute Resolution (ADR/ODR)'
https://ec.europa.eu/commission/presscorner/detail/en/IP_16_297.
supported/justified the use of ODR. In 2010, the CJEU ruled in Alassini v. Italian Telecom that the
company's online dispute resolution was fair under European treaties.68

AI Concerns
Recent years have seen increased scholarly interest in the potentials and risks of applying
Artificial Intelligence (AI) in international arbitration. Among the primary concerns raised by scholars is
the issue of bias in AI systems. A published paper from the McKinsey Global Institute explores how the
machine learning algorithms that power AI systems can inadvertently reproduce and amplify societal
biases present in the data they are trained on.69 Specifically, in the context of international arbitration,
these biases can manifest in a variety of detrimental ways, impacting the fairness of the process.

Related concerns include transparency, explainability, and accountability of AI systems. While

these systems can manage vast amounts of data and produce rapid results, their decision-making
processes are often opaque. This 'black box' phenomenon might be acceptable in certain contexts but
not in dispute resolution where decisions may significantly impact individuals and states.70

The need to ensure confidentiality in dispute resolution is another crucial point raised in the
literature. While AI can help streamline dispute resolution processes, the security of sensitive
information cannot be guaranteed.71

An additional concern cited in the literature involves the potential for AI to replace human
judgement and intuition. AI's capacity to emulate human intuition and empathy, vital for dispute
resolution, remains questionable.72 Gomez argues that AI's limitations in this regard can impact its
effectiveness in dispute resolution, particularly in complex international disputes.73

In "AI and Legal Change: Separation of Powers" Andrew C. Michaels argues that AI has the
potential to negatively disrupt the traditional separation of powers in the legal system, which is typically

68
Alassini v Telecom Italia SpA (Case C-317/08) [2010] ECR I-2213.
69
J Silberg and J Manyika, ‘Tackling Bias in Artificial Intelligence (and in Humans)’ (McKinsey & Company, 6 June
2019) <https://www.mckinsey.com/featured-insights/artificial-intelligence/tackling-bias-in-artificial-intelligence-
and-in-humans> accessed 11 September 2023.
70
J McKendrick and A Thurai, 'AI Isn’t Ready to Make Unsupervised Decisions' (2022) 15 Harvard Business Review.
71
A Majeed and SO Hwang, 'When AI meets Information Privacy: The Adversarial Role of AI in Data Sharing
Scenario' (2023).
72
C Dorsey, 'Hypothetical AI Arbitrators: A Deficiency in Empathy and Intuitive Decision-Making' (2021) 13(1)
Arbitration Law Review 12.
73
J Wu, 'Empathy in Artificial Intelligence' Forbes (17 December 2019)
<https://www.forbes.com/sites/cognitiveworld/2019/12/17/empathy-in-artificial-intelligence/?sh=7c5e0b7e6327>
accessed 11 September 2023
divided among the legislative, executive, and judicial branches.74 For instance, AI could be used to
automate the process of drafting legislation, which is traditionally the responsibility of the legislative
branch.75 This could potentially shift some power from the legislative branch to those who control the AI
systems. Michaels also cautions that the use of AI by the judicial branch to assist in decision-making
processes could potentially lead to a loss of transparency and accountability in the judicial process, as
decisions are increasingly made by algorithms rather than human judges.76

Gaps in the Literature

While the existing body of literature provides considerable insights into the application of AI in
various legal proceedings, the utilization of AI in international arbitration remains understudied when
compared to other sectors. This discrepancy presents a significant gap in knowledge.

Few studies have examined how AI can potentially alter the structure of arbitration processes,
both procedurally and substantively. The current literature primarily discusses the impact of AI on legal
practices in a generalized context, such as e-discovery, legal research, and contract analysis. However,
the specificities of international arbitration – including the complexities of cross-border legal and cultural
differences, procedural flexibility, transparency, and enforcement – make it a unique context. Therefore,
more research is required to test and measure the performance of AI in the unique aspects of
international arbitration.

There is a lack of empirical studies examining the practical implications and outcomes of using AI
in international arbitration. While some theoretical discussions propose the potential advantages and
challenges of AI's integration, empirical data that assesses the real-world consequences, both intended
and unintended, remains scarce. Such data is crucial in formulating a comprehensive understanding of
AI's role in international arbitration and its potential effects on arbitral decision-making and overall
efficiency of proceedings. Likewise, in surveying open arbitral award databases there is a clear absence
of arbitral pronouncements directly attending to the practical ramifications, boundaries, and ethical
implications of leveraging AI tools in such settings.

Additionally, ethical considerations in deploying AI in international arbitration are

underrepresented in current research. Topics such as AI's impact on the impartiality and independence

74
A C Michaels, 'Artificial Intelligence, Legal Change, and Separation of Powers' (2020) 88 U Cin L Rev 1083, 1082.
75
Ibid 1083.
76
Ibid 1085.
of arbitration, potential biases in AI decision-making, and data privacy issues are critical concerns that
need to be more extensively explored. The interplay between AI and the unique ethical dimension of
international arbitration warrants further scrutiny.

Lastly, the literature does not sufficiently address the legal implications of AI's application in
international arbitration. Questions about the legal status of decisions aided or made by AI, liability
issues, and the regulatory framework needed to govern AI's use in arbitration are areas requiring further
investigation.

3. Methodology
Research Design
This chapter presents the methodology adopted for this research on the utilization of AI in
resolving international disputes. The research for this dissertation employs a doctrinal legal research
methodology, allowing for the examination existing and theoretical uses of AI in dispute resolution in the
context of current procedures and concerns of international arbitration. The design also facilitates the
comparison of International Arbitration Laws and regulations relevant to AI across different jurisdictions,
fostering a more comprehensive understanding of the topic.

The research is designed to be exploratory and descriptive. Exploratory research provides insight
into this research problem while descriptive techniques are used to delve into specific aspects, outlining
the characteristics of AI applications in arbitration settings.

Comparative Analysis
To deepen our understanding of how AI technologies fit within the landscape of international
arbitration, a comparative analysis was undertaken specifically focusing on the technologies themselves.
Different AI tools and their functionalities, ranging from predictive analytics to natural language
processing, were studied in the context of arbitration. This methodology includes examining specific
commercial AI-powered products and platforms deployed in the legal sector. Emphasis was also placed
on projecting future advancements in AI and how they might further augment or disrupt international
arbitration proceedings. This approach provides not just a snapshot of the present landscape but also
insights into the trajectory of AI’s role in international arbitration.

Additionally, given the global nature of international arbitration, a comparative analysis is used
to study the various institutions that have adopted AI into their legal frameworks and examining how
these regulations accommodate the intricacies of arbitration. This comparative analysis not only aids in
understanding the extent to which AI is being used in different regions but also in highlighting best
practices and potential pitfalls in the integration of AI into international arbitration.

Data Collection Methods

The research depends entirely on secondary data gathered from various sources, including
academic journals, legal databases, AI dispute resolution platform data, national laws and legislative
documents, International Legal Instruments, and where possible, arbitral awards, case law. The
examination of secondary data allows for a detailed study of the practical and theoretical aspects of AI in
dispute resolution.

Given the rapid advancement of AI and its applications in arbitration, documents from the last
decade are the focus to ensure the study's relevance and applicability. Grey literature like reports and
articles from reliable institutions are also included to supplement academic literature.

Data Analysis
The collected data is scrutinized through a doctrinal analytical procedure. This approach involves
an in-depth examination and critical review of legal doctrines, statutes, principles, and case law related
to the use of AI in international arbitration. The doctrinal analysis method entails a comprehensive study
of the existing legal frameworks, jurisprudential developments, and scholarly commentaries, focusing on
the legitimacy, effectiveness, and implications of AI applications in international arbitration contexts.

The doctrinal analysis aims to uncover underlying principles, identify inconsistencies, and gauge
the adequacy and effectiveness of legal norms and institutional practices concerning AI in international
arbitration. The primary focus will be to assess the coherence and evolution of AI use as well as changing
legal doctrines to discern any emerging patterns or trends within the legal frameworks, jurisprudence,
and practice. This qualitative analysis method will be instrumental in understanding, interpreting, and
forming a reasoned opinion on the legal aspects and ramifications of employing AI in international
arbitration processes.

The detailed examination of legal texts, judicial opinions, and scholarly works enables a profound
understanding of the advantages, disadvantages, and effects of using AI in international arbitration,
laying the foundation for informed discussions, interpretations, recommendations, and conclusions on
the subject matter.
Ethics, Trustworthiness and Credibility of the Research
Although this research does not involve human subjects, ethical considerations remain relevant,
particularly concerning the accurate representation and interpretation of secondary data. The research
incorporates several strategies to ensure its credibility, including thorough source validation, the use of
multiple data sources for triangulation, and a transparent, replicable analytical process.77

Limitations
Given the growing and evolving nature of AI, potential limitations of the study include the
possibility of incomplete or inaccessible data and potential disparities in data quality across sources and
jurisdictions. Also, the use of English language sources might limit the scope to primarily English-
speaking jurisdictions. However, through careful source selection and the inclusion of multiple data
sources, these limitations are mitigated.

4. Findings
The objective of this research has been fourfold: to investigate the current applications of AI in
international arbitration, to assess its potential to further alleviate court burden, to identify the ethical
and legal challenges, and lastly to provide recommendations to address these challenges.

The Findings begin with examining the growing interplay between AI technologies and
international arbitration in contemporary dispute resolution processes.

The Findings then analyze the benefits offered by AI in streamlining international arbitration and
improving its efficiency, predictability, and effectiveness. Next, building on present applications, the
findings explore the Future AI and International Arbitration as well as current and fundamental
limitations of AI, drawing attention to the potential drawbacks and areas of improvement.

The examination continues with Ethical and Legal Challenges Associated with AI in International
arbitration, dissecting the moral and legal predicaments brought about by the application of AI in
international arbitration.

Lastly, the Summary of Findings offers an all-encompassing review of the investigation's

discoveries, synthesizing the findings and establishing a coherent narrative.

77
Y Lincoln and E Guba, 'Naturalistic Inquiry' (SAGE Publications, 1985).
AI Advancements in International Arbitration
The findings have revealed transformations in international arbitration paradigms, largely
propelled by technological innovations. These include a rise in virtual arbitration proceedings, enhanced
rules for Evidence and Disclosure, the selection and role of arbitrators, and the enforcement of awards.

Virtual Arbitration Meetings: The research findings have shown that virtual arbitration hearings
stand as a viable and innovative substitute for traditional in-person arbitration hearings. Notably, AI
technologies have emerged as pivotal components, amplifying the efficacy, security, and accessibility of
virtual arbitration processes, thereby revealing newfound potentials in the realm of dispute resolution.78

The International Chamber of Commerce (ICC) has provided guidelines since 2004 on the use of
new technologies in arbitration. In 2017, the ICC introduced rules for small claims cases where
arbitrators could base their awards solely on documents (aided by technology analysis), without the
need for hearings. In 2022, it provided further guidance on the use of technologies for virtual hearings.79
Additionally, WIPO Center, JAMS, AAA-ICDR, and the International Institute for Conflict Prevention &
Resolution (CPR) allow hearings via videoconference. AI-driven platforms like Zoom and Microsoft Teams,
or specialized legal conference tools, offer intelligent scheduling, speech and face recognition, enhanced
security features, and advanced interaction capabilities, ensuring a smooth and secure environment for
conducting hearings. Of note, the recently released Zoom AI Companion can enhance the efficiency and
clarity of virtual arbitration hearings by enabling participants to create summarizations of hearings in
real-time and streamline communication with intelligent chat and email composition. Additionally, with
Smart Recording, parties can easily revisit critical parts of hearings through organized chapters and
highlighted segments.80

In support of virtual hearings, the case Hanaro Shipping v. Cofftea Trading, found that there was
no procedural imbalance when one party's witnesses gave evidence in person while the other party's
witnesses gave evidence only by video link.81

In the case of Eli Lilly and Company v. Government of Canada, which primarily dealt with patent
rights under NAFTA, the arbitration court made available documents, submissions, as well as a live

78
M Moscati and others, Comparative Dispute Resolution (Edward Elgar 2020) 424.
79
ICC, 'Leveraging Technology for Fair, Effective and Efficient International Arbitration Proceedings' (ICC Publication,
18 February 2022).
80
(Zoom AI Companion, 5 October 2023) <https://explore.zoom.us/en/ai-assistant/> accessed 5 October 2023.
81
Hanaro Shipping v Cofftea Trading [2015] EWHC 4293 (Comm) [16].
webcast of the hearing.82 In such cases, secure online AI-powered document-sharing platforms facilitate
encrypted and secure transmission of documents, ensuring integrity and confidentiality of sensitive
information.83

In response to the Covid-19 Pandemic, arbitration institutions further adapted to the new reality
where personal contact posed a risk of infection. They expanded virtual hearings, signing decisions and
awards electronically, and began integrating supporting AI-based technologies into all aspects of the
arbitration procedure.84 Between March and April 2020, nearly all arbitration institutions updated their
rules to accommodate the possibility of conducting the entire procedure remotely. Institutions like ICC,
SIAC, HKIAC, and LCIA confirmed their continued operations and the management of new and existing
cases.85 The ICC, for example, released a "Guidance Note on Possible Measures Aimed at Mitigating the
Effects of the Covid-19 Pandemic" in April 2020.86 This note provided a list of measures that arbitrators
could adopt without risking challenges from any of the involved parties. Latin American institutions, such
as the Arbitration and Mediation Center of the Santiago Chamber of Commerce (CAM), also encouraged
the use of technological platforms for hearings.

Governments also played a role in promoting these technological shifts. In Colombia, the
government issued Decree No. 491 (2020), which mandated the continuation of arbitration, conciliation,
and other alternative dispute resolution mechanisms, barring any insurmountable technical, personal, or
legal challenges.87

In the case of Sky Power v. IrAero88, decided under the London Court of International Arbitration
(LCIA), the parties initially agreed to a semi-virtual hearing due to the Covid-19 Pandemic. However, with
Sky Power's key witness unable to travel to Moscow for health concerns, they suggested a fully virtual

82
Eli Lilly and Company v The Government of Canada, UNCITRAL, ICSID Case No UNCT/14/2 (16 March 2017).
83
'Secure AI training: When on-premise beats the cloud' (Equus Compute Solutions, 2023)
<https://www.equuscs.com/secure-ai-training/> accessed 04 September 2023.
84
T Landon and K von der Weid, 'The Impact of COVID-19 on International Arbitration Procedure' in T Landon and K
von der Weid (eds), The Impact of Covid on International Disputes (2022)
85
Ibid 53.
86
ICC, 'ICC Guidance Note on Possible Measures Aimed at Mitigating the Effects of the COVID-19 Pandemic', ICC
Dispute Resolution Bulletin, 2020(2), 1-10 (2020).
87
H Grigera Naón and B Arp, ‘Virtual Arbitration in Viral Times: The Impact of Covid-19 on the Practice of
International Commercial Arbitration’ (American University Washington College of Law, 23 May 2020)
<https://www.wcl.american.edu/impact/initiatives-programs/international/news/virtual-arbitration-in-viral-times-
the-impact-of-covid-19-on-the-practice-of-international-commercial-arbitration/> accessed 7 September 2023.
88
Sky Power v Iraero (LCIA, [2022]).
hearing—a motion opposed by Iraero Airlines. The arbitrator, assessing the situation, directed a fully
virtual hearing.

Upon the arbitrator's favorable decision for Sky Power, enforcement of the award was sought in
Hong Kong. Iraero Airlines contested, arguing the virtual hearing diverged from the original agreement
and compromised its presentation. In Sky Power Construction Engineering Limited v Iraero Airlines JSC
[2023] HKCFI 155889, the Hong Kong Court of First Instance determined that an objection to a fully virtual
hearing is insubstantial grounds to resist the enforcement of an arbitration award. The court
underscored the pandemic-induced acceptability of remote hearings and upheld that arbitrators hold
discretion in case management, unbounded by prior agreements on semi-virtual hearings. This ruling
offers guidance on future arbitration disputes concerning virtual hearing modalities.

Enhanced Rules for Evidence and Disclosure through AI-Powered Technologies: The
modifications to the rules for evidence and disclosure by international arbitration bodies showcase an
emphasis on enhanced efficiency, transparency, and procedural flexibility. AI-powered technologies are
instrumental in supporting these modifications, providing tools and solutions that cater to the evolving
needs and complexities of international arbitration. The incorporation of AI not only aligns with the
advancements in arbitration procedures but also propels the international arbitration mechanism into a
new era of technological integration and sophistication.

The ICC introduced revisions to its Arbitration Rules effective from 1 January 2021,90 with
emphasis on case management and expeditious dispute resolution procedures. Machine learning and
natural language processing, can aid in expeditiously analyzing and managing cases, by processing
extensive information, predicting outcomes and identifying patterns. Enhanced transparency in
evidentiary disclosures and flexibility in dealing with documentary evidence were introduced, allowing
tribunals to decide the level of scrutiny applied to each piece of evidence. AI-powered document
analysis tools can support these innovations by automating the scrutiny of documents and identifying
relevant information swiftly and accurately.

The LCIA amended its Arbitration Rules effective from 1 October 2020.91 The rules accentuated
efficiency in the arbitration process, specifically stating that the Arbitral Tribunal shall have the discretion

89
Sky Power v Iraero (HKCFI, [2023]).
90
ICC, ‘ICC Arbitration Rules 2021’ (International Chamber of Commerce 2021) <www.iccwbo.org>.
91
LCIA, ‘LCIA Arbitration Rules 2020’ (London Court of International Arbitration 2020) <www.lcia.org>.
to decide the admissibility, relevance, materiality, and weight of the evidence.92 AI can significantly
support such discretionary powers by providing advanced analytics, insights and predictive coding,
facilitating a more informed and efficient decision-making process on the relevance and materiality of
the evidence, thus endorsing a more proactive approach in evidence evaluation.

The ICDR’s March 1, 2021 amendments,93 concentrate on the integrity, efficiency, and practicality
of international arbitration processes. AI-powered technologies can contribute by ensuring data integrity
and facilitating the secure and efficient exchange of information. The newly instituted rules reflect on the
disclosure of documents, taking a more considered approach to the exchange of information and
presenting clearer guidelines regarding document production. AI can support this approach by
streamlining document review processes, ensuring compliance with disclosure rules, and reducing the
time and resources required to comply with the new disclosure norms.

Additionally, it should be noted that the widespread adoption of the IBA Rules on the Taking of
Evidence in International Arbitration has been amplified by AI's ability to process and analyze vast
amounts of data.94

The case of BSG Resources Limited, Vale S.A. v. Republic of Guinea (ICSID Case No. ARB/14/22)95
is an example of how AI-powered e-discovery tools were significantly leveraged to handle a substantial
volume of evidence and disclosure documents.

The case revolved around a dispute concerning a mining concession in Guinea, involving
allegations of corruption related to the acquisition of mining rights. This arbitration procedure required
the examination of an extensive amount of documents and electronic data, which led the tribunal and
the parties involved to adopt sophisticated e-discovery software. These tools were instrumental in
searching, filtering, and organizing the extensive digital data and documentation, which were disclosed
by the parties, allowing the tribunal to manage the evidentiary material efficiently.96

The application of advanced technology and e-discovery tools enhanced the accuracy and
efficiency of the evidence review process, allowing relevant and material evidence to be highlighted and

92
ICDR, ‘ICDR Arbitration Rules 2021’ (International Centre for Dispute Resolution 2021) <www.icdr.org>
93
Ibid.
94
International Bar Association, 'Rules on the Taking of Evidence in International Arbitration' (17 December 2020)
<https://www.ibanet.org/MediaHandler?id=def0807b-9fec-43ef-b624-f2cb2af7cf7b> (accessed 7 July 2023).
95
BSG Resources Limited, Vale S.A. v Republic of Guinea (ICSID, [2022]).
96
Ibid,
scrutinized meticulously. The transparency and integrity of the evidence and disclosure process were
thus substantially bolstered, establishing a precedent for leveraging advanced technologies in
international arbitration procedures.

Use of AI in the Selection of Arbitrators: An initial step in the arbitration process is the often-
laborious task of selecting and appointing arbitrators who bear the responsibility to adjudicate disputes.
This phase requires parties to invest significant time and resources, which, with the assistance of AI, can
be significantly streamlined. Recent advancements have seen the development of several platforms
aimed at aiding this crucial phase of arbitration.

Access to tribunal and arbitrator data is pivotal. To make informed decisions, access to
comprehensive data regarding individual arbitrators, their past experiences, perspectives on various
issues, current case load, and procedural practice, is vital. Various arbitral institutions like the ICDR, ICC,
and LCIA have made available data about potential arbitrators, as well as cost and duration of
proceedings.97

Programs such as the International Commercial Arbitration Toolkit are collecting qualitative and
quantitative data on professionals and cases, focusing on elements critical for the selection of
arbitrators. Using AI algorithms, these platforms optimize the selection process by analyzing extensive
data pertaining to the history of potential candidates and their relevance to each specific case.98 This
contributes to an enhanced understanding and linkage between the potential arbitrators and the
intrinsic nature of the disputes at hand.

These tools bring forth an impartial and objective methodology to the selection process,
minimizing human biases and subjectivity. They allow for selections to be grounded on facts, objective
criteria, and data analytics, ensuring arbitrators are not overburdened and pinpointing potential conflicts
of interest. AI provides equitable grounds for comparison between novice and veteran arbitrators,
bypassing opacity in the process and preventing the recurrent appointment of the ‘usual faces,’ thereby
opening avenues for diverse and potentially less seasoned candidates.

97
J Commission, 'The increasing use of data analytics in international arbitration' (2020) New York Law Journal
<https://www.law.com/newyorklawjournal/2020/11/20/the-increasing-use-of-data-analytics-in-international-
arbitration/> accessed 07 September 2023.
98
International Commercial Arbitration Toolkit, 'International Commercial Arbitration (ICA) tracker' (2023) <
https://www.dentons.com/en/about-dentons/news-events-and-awards/news/2023/april/dentons-launches-
innovative-international-arbitration-user-toolkit > accessed 07 September 2023.
 The deployment of AI tools can also be used to encourage diversity, integrating factors like
gender, race, age, geography, language, and ethnicity into models, thereby potentially narrowing the
diversity gap.

Tech-Enhanced Enforcement of Arbitral Awards: The enforcement of International Arbitration

awards is enhanced by AI tools enabling rapid verification of awards through advanced search
capabilities of relevant databases (awards, laws, associated documents).99 NLP tools also enhance
enforcement by facilitating secure, confidential communication between relevant entities during the
award process. Likewise, electronic signature and document verification tools streamline the signing and
authentication processes, safeguarding the documents' authenticity.100 These technologies also foster
international cooperation, enhancing the cross-border enforcement of awards. This is achieved by
improving coordination between international agencies and facilitating secure cross-country information
exchange.

ODR platforms are now offering swifter, more accessible dispute resolution mechanisms using AI,
ensuring awards align with relevant international laws, treaties, and conventions. Combining these
platforms with blockchain technology can create "smart award agreements" automating settlements and
ensuring adherence to the awarded decisions. This automation notably decreases time and costs in
international arbitration, making it more appealing for parties in cross-border disagreements.101

The 2021 UK Law Commission noted that smart legal contracts can vary – from plainly worded
contracts capable of auto-execution to those purely in computer code.102 Integrating arbitration
agreements in these contracts can be simple or complex, especially when coded. The implications of
such coded agreements on the binding nature under English law and their alignment with the New York
Convention are yet to be fully elucidated.

However, the use of a turnkey AI-powered arbitration system can be examined through the Hong
Kong’s Electronic Business-Related Arbitration and Mediation Platform (eBRAM).103 This platform utilizes
AI functions including text translation, transcription, and advanced user authentication technologies,

99
T Landon and K von der Weid, 'The Impact of COVID-19 on International Arbitration Procedure' in T Landon and K
von der Weid (eds), The Impact of Covid on International Disputes (2022) 263.
100
M S Karim, 'Artificial Intelligence: An Undiscovered Future of Arbitration' (2019) 22(2) Int A.L.R. 47.
101
T Landon and K von der Weid, 'The Impact of COVID-19 on International Arbitration Procedure' in T Landon and
K von der Weid (eds), The Impact of Covid on International Disputes (2022) 116.
102
Law Commission, Smart legal contracts (Law Com No 401, 2021).
103
eBRAM International Online Dispute Resolution Centre <https://www.ebram.org/services.html> accessed 22
July 2023.
aimed at facilitating the entire arbitration processes. eBRAM offers conclusive, binding, and enforceable
awards and is adept at handling both domestic and international disputes.104

General Advantages of AI in International Arbitration

There are several key, broad-ranging benefits that AI brings to International Arbitration. Most
notably AI clearly enhances efficiency and speed, ensuring the expeditious resolution of disputes.
Additionally, AI improves accuracy and consistency, removing human errors and biases, thereby fostering
trust. It also can be cost-effective by minimizing human expenditure and lastly, it can augment
accessibility and transparency in international arbitration, opening doors for a wider pool of participants
and creating a more level playing field. The exploration of these four significant advantages offers insight
into the transformative potential of AI in International Arbitration, underscoring its role in shaping the
future of international dispute management.

Efficiency and Speed: AI significantly boosts efficiency in document review during arbitration. AI,
through technologies like Machine Learning and Natural Language Processing, can analyze thousands of
documents in a fraction of the time it would take a human team, without the risk of fatigue or decreased
productivity over time.105 Additionally, as noted, AI predictive analytics can utilize machine learning
algorithms to anticipate the outcome of legal cases. AI-based solutions such as Thomson Reuters’ HIGHQ
Case Management solution save considerable time and resources by providing an early, statistically
backed assessment of case details, strengths, and weaknesses.106 Lastly, AI-based solutions such as
Practice Panther’s Law Practice Management Software can automate various administrative tasks like
scheduling, billing, and correspondence, which are integral to arbitration but time-consuming.107

Accuracy and Consistency: AI's automated systems can help minimize the potential for human
error, a critical consideration in international arbitration where high-stakes decisions often hinge on
meticulous document review and complex analyses. By taking over tasks that can lead to human error, AI
tools such as SmartAdvocate can contribute to more accurate arbitration processes and outcomes.108

104
Ibid.
105
A Schmitz and C Rule, 'OArb Enters the Age of Artificial Intelligence' (2023) 29 Dispute Resolution Magazine 36
<https://www.proquest.com/central/docview/2841558523/45CBB453357D48FCPQ/1?accountid=14089> accessed
29 August 2023.
106
Thomson Reuters, 'HIGHQ Litigation Management'
<https://legal.thomsonreuters.com/en/products/highq/litigation-management> accessed 6 August 2023.
107
PracticePanther, 'Law Practice Management Software' <https://www.practicepanther.com/> accessed 6 August
2023.
108
SmartAdvocate, 'Legal Case Management Software' <https://www.smartadvocate.com/features/revolutionizing-
case-management-with-ai-technology/> accessed 6 August 2023.
Additionally, AI's ability to analyze and learn from historical case data can help ensure more consistent
decision-making in arbitration.

Cost-Effectiveness: The automation capabilities of AI can significantly reduce the man-hours

required for case preparation and administration, leading to considerable cost savings. Automated
document review and administrative task handling can reduce the time legal teams need to spend on
these tasks, allowing them to focus on more complex, strategic elements of arbitration.109 Likewise, by
automating various aspects of case management, AI can lower the overall costs of arbitration. In
addition to reducing man-hours, AI's efficiency and speed can lead to shorter arbitration timelines,
further reducing costs.

Accessibility and Clarity: As described above, AI is effectively facilitating remote arbitration,

particularly important in the context of international disputes. AI technologies, including blockchain and
smart contracts, can facilitate the execution and enforcement of arbitral awards remotely. AI can also
contribute to decision clarity in arbitration by providing data-driven insights. The use of AI in tracking
and analyzing arbitration trends and outcomes can offer valuable insights for parties, counsels, and
arbitrators, enhancing the clarity and predictability of the process.110

Limitations of AI in International Arbitration

Despite the various advantages and future potential discussed in this paper, the application of AI
in international arbitration is not without its challenges, which this section will explore in detail.

Grasping Complex Legal Concepts: Although AI has shown promise in understanding and
analyzing vast quantities of data, its ability to grasp complex legal concepts is currently limited. Legal
reasoning involves an intricate balance of interpreting statutory text, precedent, and the principles of
justice and equity, often within a specific cultural and historical context. Such interpretation often
involves the consideration of a wide range of factors beyond the literal text, which may prove challenging
for AI to appreciate fully.111 In international arbitration, this becomes even more complex due to the
involvement of different legal systems and cultures.

109
P B Marrow, M Karol and S Kuyan, 'Artificial Intelligence and Arbitration: The Computer as an Arbitrator-Are We
There Yet?' (2020) 74 Dispute Resolution Journal 35.
110
G L Benton and S K Andersen, 'Technology Arbitration Revisited' (2020) 74 Dispute Resolution Journal 1.
111
Y Bathaee, 'The Artificial Intelligence Black Box and the Failure of Intent and Causation' (2017) 31 Harv JL & Tech
889.
 Understanding Nuanced Human Interaction and Emotions: Another challenge in applying AI to
international arbitration pertains to its limitation in understanding nuanced human interaction and
emotions. At the heart of many international disputes are complex human relationships and emotions,
which play a critical role in resolution and settlement processes. However, AI's ability to understand,
interpret, and respond to human emotions is currently rudimentary, often producing unsatisfactory or
inappropriate results.112 Moreover, language nuances, cultural idiosyncrasies and the non-verbal
elements of human communication may be missed or misinterpreted by AI.

Decision Interpretability and Emotional Intelligence: Finally, a significant challenge for AI in

international arbitration is the issue of decision interpretability and generalizability. While AI models,
especially those based on machine learning, can generate predictions or decisions, understanding the
reasoning behind these decisions often remains difficult. This lack of transparency can lead to mistrust
and reluctance to adopt AI technologies.113 Additionally, the AI models' capacity to generalize from the
learned patterns to new, unseen situations is also limited. For AI to be used effectively in dispute
resolution, it needs to demonstrate reliable performance across a wide variety of cases and jurisdictions,
many of which can be highly unique and complex.

The application of AI in international arbitration raises pertinent questions about its ability to
fully replicate or substitute human intelligence, judgement, and ethical considerations.114 For instance, as
of the writing of this paper, AI lacks emotional intelligence, which is fundamental to the roles of lawyers,
arbitrators, and judges, especially in matters that require the interpretation of the law or the assessment
of parties' intentions.115

Data Access and Security: Furthermore, AI systems are only as good as the data they are trained
on. Hence, they are susceptible to the GIGO (Garbage In, Garbage Out) principle.116 AI’s predictions and
analyses are often based on previous patterns and outcomes. Therefore, a novel case might prove
problematic for AI to handle.

112
J McKendrick and A Thurai, 'AI Isn’t Ready to Make Unsupervised Decisions' (2022) 15 Harvard Business Review.
113
Ibid
114
C Dorsey, 'Hypothetical AI Arbitrators: A Deficiency in Empathy and Intuitive Decision-Making' (2021) 13(1)
Arbitration Law Review 12.
115
O Brookhouse, 'Can Artificial Intelligence understand emotions?' (Telefonica, 17 April 2020)
<https://business.blogthinkbig.com/can-artificial-intelligence-understand-emotions/>.
116
R Ozminkowski, 'Garbage in, Garbage Out' (Medium, 2021) <https://towardsdatascience.com/garbage-in-
garbage-out-721b5b299bc1> accessed 01 September 2023.
 The training and operation of AI in international arbitration also faces considerable challenges
due to the nature of the data involved. As mentioned, data related to dispute resolution is privileged or
confidential, thereby making it inaccessible for the training of AI. Since AI systems often require
extensive data to generate reliable predictions and insights, the comparative paucity of available data
could significantly limit the utility and accuracy of AI output.117

In addition to data scarcity and privilege, there are also issues surrounding data security and
privacy. Misuse of data, either intentionally or unintentionally, can lead to breaches of privacy and
confidentiality, potentially attracting legal penalties.118

Existing Laws and Regulations: In reviewing the landscape of international arbitration rules set
by notable institutions, alongside national arbitration laws, a discernible recognition of the evolving role
of technology emerges. Many contemporary regulations, while not addressing AI explicitly, implicitly
permit its utilization for facilitative and auxiliary functions such as evidence assessment, data analytics,
or procedural management. Such provisions may be read to appreciate the efficiency and accuracy AI
tools bring to these non-decisive processes. However, when we consider the core duties of the
arbitrator, several existing laws may prohibit the use of AI or give grounds for challenges to judgements
and awards.

If an arbitration award is based on procedures that contravene the arbitration laws of the
country where enforcement is sought or of the country agreed upon as the place of arbitration, the party
against whom the award is invoked might have legitimate grounds to challenge or resist the enforcement
of that award.

In certain jurisdictions, the enforcement of an award determined by or with the assistance of

artificial intelligence is not permissible. As such, to avoid potential nullification, parties and arbitrators
might be hesitant to utilize AI. For awards to be enforced in countries like France119 or Peru120, a human
arbitrator is essential. Furthermore, the Arbitration Acts of nations such as Brazil121, the Netherlands122,

117
C M de Westgaver, 'Canvassing views on AI in IA: The Rise of Machine Learning' (Kluwer Arbitration Blog, 12 July
2023) <https://arbitrationblog.kluwerarbitration.com/2023/07/12/canvassing-views-on-ai-in-ia-the-rise-of-
machine-learning/> accessed 17 August 2023.
118
'Secure AI training: When on-premise beats the cloud' (Equus Compute Solutions, 2023)
<https://www.equuscs.com/secure-ai-training/> accessed 04 September 2023.
119
French Code of Civil Procedure, Art 1450
120
Arbitration Act of Peru LEY Nº 1017 Art 20.
121
Arbitration Act of Brazil LEI Nº 9.307; Art 10.
122
Netherlands Arbitration Act Nº1450; Art 1023 Rv.
Ecuador123, and Colombia124 specifically designate arbitrators as 'human' or mandate their individual
action.

Future AI and International Arbitration

Emotion AI: Emotion AI, is in the pursuit of deciphering, construing, synthesizing, and replicating
human emotions with an unparalleled precision and depth. It strives to develop systems and devices that
can understand, interact with, and respond to human emotions in a more personalized and adaptive
manner. If advanced, Emotion AI will be able to substitute human reasoning, which is crucial for evaluating
the credibility of witnesses and evidence. Assessing the reliability of presented evidence, the sincerity of
witnesses, and the overall believability of a case is a complex task that requires a deep understanding of
human behavior and intention.125

Applications of emotion AI are found in various fields like marketing, healthcare, and particularly
law enforcement.126 From detecting deceit during interrogations to identifying potential threats based on
emotional distress, the technology promises improved accuracy and efficiency. If trained properly, these
technologies are transferable to international arbitration where regional/cultural differences can
interfere effective arbitration engagements.

Emotion AI could be instrumental in modeling human reasoning by providing insights into the
emotional states and cognitive processes of the disputing parties. This can be valuable in understanding
the underlying motivations and intentions, thereby allowing for more informed and empathetic
decisions. Additionally, Emotion AI can play a crucial role in navigating cultural nuances and differences
in international arbitration, ensuring that each party’s sentiments, values, and concerns are accurately
interpreted and addressed. This can reduce misunderstandings and facilitate smoother communication
between parties from diverse cultural backgrounds. Moreover, by simulating empathy, Emotion AI can
foster an environment of mutual understanding and respect. This enables arbitrators to adapt their

123
Arbitration Act of Ecuador LEY Nº 14-417; Art 19.
124
Colombian Arbitration Act LEY Nº 1563; Art. 7
125
M Somers, 'Emotion AI, Explained' (MIT Sloan School of Management, 8 March 2019)
<https://mitsloan.mit.edu/ideas-made-to-matter/emotion-ai-explained>; see also Pickard (n 82).
126
K Roemmich, F Schaub and N Andalibi, 'Emotion AI at Work: Implications for Workplace Surveillance, Emotional
Labor, and Emotional Privacy' (2023) Proceedings of the 2023 CHI Conference on Human Factors in Computing
Systems 1-20.
communication and approach, aligning with the emotional and cognitive states of the disputants, which
can contribute to more effective resolution of conflicts.127

Artificial General Intelligence (AGI): Artificial General Intelligence (AGI), an entity not yet
brought into existence, is envisaged to possess the capability to comprehend, acquire, and apply
knowledge across a plethora of tasks, distinguishing it substantially from existing Narrow Artificial
Intelligence (AI).128 While AI has found its utility in legal realms—primarily in machine learning, natural
language processing, and expert systems—to facilitate tasks like legal research, contract analysis, and
discovery, AGI is foreseen to surpass these functionalities, embracing advanced cognitive tasks such as
strategizing, negotiation, and advising, traditionally the exclusive domain of human legal practitioners.129

AGI, with advanced cognitive abilities, will surpass narrow AI in processing extensive legal
databases and case histories, promising expedited and economical legal research. It will pinpoint and
correlate pertinent legal precedents with heightened accuracy and context comprehension, elevating the
quality of legal counsel provided.

Additionally, AGI’s capability to assimilate knowledge across domains will illuminate risks in
contract analysis by interpreting contractual obligations within the broader commercial and regulatory
frameworks. Its learning from diverse previous contract drafting and review endeavors will yield superior
automation, mitigating errors and enhancing efficiency.

Likewise, AGI’s holistic understanding will facilitate accurate prediction of case outcomes and
formulation of astute litigation strategies by analyzing a broader spectrum of factors potentially
overlooked by narrow AI. Its risk understanding and prediction will augment decision-making and
compliance adherence amidst evolving regulations.

Lastly, AGI will efficiently process, analyze, and organize voluminous data, alleviating the
workload on human practitioners. Its exceptional task automation and intelligent data analysis will
hasten litigation processes. Envisioning AGI as a co-mediator in future legal arbitrations appears

127
R Srinivasan and BSM González, ‘The Role of Empathy for Artificial Intelligence Accountability’ (2022) 9 Journal
of Responsible Technology 100021.
128
S Rayhan, 'Ethical Implications of Creating AGI: Impact on Human Society, Privacy, and Power Dynamics' (2023)
Artificial Intelligence Review.
129
A Aidid and B Alarie, The Legal Singularity: How Artificial Intelligence Can Make Law Radically Better (University
of Toronto Press 2023).
plausible, where it will serve alongside human mediators, transcending the assistant role currently
attributed to AI.

Ethical and Legal Challenges Associated with AI in International Arbitration

AI Ethics are a set of values, principles, and techniques that employ widely accepted standards of
right and wrong to guide moral conduct in the development and use of AI technologies.

Infographic: AI Values and Principles

AI Bias and the Implications for Fairness

While AI promises to increase efficiency, its application raises significant ethical and legal issues,
most notably the problem of bias. AI bias refers to systematic errors in outputs produced by machine
learning models, which create unfair and prejudiced outcomes. These prejudices typically mirror those
existing in society, such as discrimination based on gender, race, age, or nationality.130

Bias in AI systems could compromise the impartiality of international arbitration mechanisms. AI

algorithms are often based on data which may contain biases, resulting in decisions that are prejudiced

130
O Osoba & W Welser IV, 'An Intelligence in Our Image – The Risk of Bias and Errors in Artificial Intelligence' (Rand
2017) 5.
or discriminatory.131 For instance, machine learning algorithms trained on biased data may perpetuate or
even amplify existing social biases.132

Unchecked bias can undermine fairness and impartiality, raising serious concerns about the use
of AI in critical decision-making processes. Mitigating these risks requires more representative data
collection, careful algorithmic design, and ongoing scrutiny of AI systems' outcomes.

Bias in AI can infiltrate the system in several ways. The most common is through biased training
data. If the data used to train the AI system is skewed, incomplete, or discriminatory, the AI system will
replicate those biases. The AI's algorithms and design can also inadvertently introduce bias.

In international arbitration, the principles of fairness and neutrality play a significant role. The
ideal for a dispute resolution process is that it must be free from any sort of bias and prejudice to ensure
just outcomes.133 The United Nations Commission on International Trade Law (UNCITRAL) Model Law on
International Commercial Arbitration underscores the importance of an impartial and independent
arbitral tribunal.134

There are several well-documented instances of bias in AI systems. One of the most infamous
cases involved Google's facial recognition technology which misidentified black faces as gorillas.135

In another case, Amazon's AI recruitment tool displayed gender bias, favoring male candidates
over females. The system was trained on resumes submitted to Amazon over a ten-year period, the
majority of which came from men.136

Predictive policing is another domain where AI bias is evident. In the US, an AI system known as
COMPAS used to predict recidivism was found to have racial bias, with a higher false-positive rate for

131
R Susskind, Tomorrow's Lawyers: An Introduction to Your Future (Oxford University Press 2013) 45.
132
O Osoba & W Welser IV, 'An Intelligence in Our Image – The Risk of Bias and Errors in Artificial Intelligence' (Rand
2017) 13.
133
D Carneiro, P Novais, F Andrade, J Zeleznikow & J Neves, 'Online dispute resolution: an artificial intelligence
perspective' (2014) 41 Artificial Intelligence Review 211-240.
134
UNCITRAL, 'Model Law on International Commercial Arbitration 1985'
<https://uncitral.un.org/en/texts/arbitration/modellaw/commercial_arbitration> accessed 30 July 2023.
135
G N and K Hill, ‘Google’s Photo App Still Can’t Find Gorillas. and Neither Can Apple’s.’ (The New York Times, 22
May 2023) <https://www.nytimes.com/2023/05/22/technology/ai-photo-labels-google-apple.html> accessed 11
August 2023.
136
Dastin J, ‘Amazon Scraps Secret AI Recruiting Tool That Showed Bias against Women’ (Reuters, 10 October 2018)
<https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G> accessed 11 August
2023.
black defendants than for white defendants.137 Likewise, in the UK, the Harm Risk Assessment Tool
(HART system) was found to exhibit AI bias, especially in its predictions related to individuals' likelihood
to commit crimes. Studies indicated that the system disproportionately misclassified individuals based
on factors like race, leading to concerns about its fairness and accuracy in real-world applications.138

Infographic: Various forms of AI bias

Such instances of bias in AI systems highlight the critical need for neutrality and impartiality
particularly in international arbitration, where unbiased decision-making is paramount to ensuring just
and equitable outcomes for all parties involved.

As discussed, a primary issue is access to data so hypothetically, an AI-powered arbitration

system that favors entities from specific jurisdictions due to the disproportionate number of arbitration
cases present in the training data may result in the AI system being more likely to predict outcomes
favorable to those entities, which raises serious fairness and impartiality concerns.

137
S Corbett-Davies and others, ‘A Computer Program Used for Bail and Sentencing Decisions Was Labeled Biased
against Blacks. It’s Actually Not That Clear.’ (The Washington Post, 17 October 2016)
<https://www.washingtonpost.com/news/monkey-cage/wp/2016/10/17/can-an-algorithm-be-racist-our-analysis-
is-more-cautious-than-propublicas/> accessed 11 August 2023; Loomis v State (Wis 2016) 881 N.W.2d 749, 767.
138
M Burgess, 'UK Police are Using AI to Inform Custodial Decisions—But It Could be Discriminating Against the
Poor' The Wired (3 January 2018) <https://www.wired.co.uk/article/police-ai-uk-durham-hart-checkpoint-
algorithm-> accessed 11 August 2023.
 Likewise, an AI-powered arbitration system that is trained on arbitration awards and contract
data predominantly from Western countries may lead to cultural bias in the review strategies proposed
by the system.

The potential bias in AI threatens not only the fairness of the arbitration process but also its
acceptance by the parties involved. The perception of an unjust process may undermine the legitimacy
of arbitration decisions and impede the enforcement of arbitral awards.139 Furthermore, the opaqueness
often associated with AI's decision-making process can make it difficult for parties to understand and
challenge unfair decisions, thus violating the principle of transparency.140

Given these considerations, further research and regulations are needed to address AI bias in
international arbitration, ensuring the continued applicability of principles of fairness and neutrality.

AI's Limitations in Recognizing Cultural Differences

While AI can analyze patterns and derive conclusions based on data, it struggles to grasp the
cultural nuances and issues inherent in human communication. This is significant in international
arbitration where cultural sensitivity is pivotal - given the varying cultural backgrounds, beliefs, and
norms that influence the proceedings. Ignoring or misunderstanding these can lead to miscarriages of
justice or unsatisfactory resolutions.

Shortcomings of AI in Addressing Cultural Aspects:

• Communication Styles: AI systems are designed to process information based on pre-determined

algorithms and parameters. Different cultures have distinct styles of communication – direct
versus indirect, high-context versus low-context. For example, in many Eastern cultures, indirect
communication is favored to prevent confrontation, while Western cultures might prioritize
directness. An AI system might misinterpret indirectness as ambiguity or lack of clarity, not
grasping the cultural nuance behind the approach.
• Importance of Relationships and Face-Saving: Many AI systems are not designed to discern the
importance of maintaining face or honor pivotal in many Asian cultures. Actions that might be
seen by an AI as straightforward might, in a human cultural context, be seen as affronts or

139
UNCITRAL, 'Convention on the Recognition and Enforcement of Foreign Arbitral Awards' (New York, 1958) art III
<https://uncitral.un.org/en/texts/arbitration/conventions/foreign_arbitral_awards> accessed 30 July 2023.
140
M Scherer, 'Artificial Intelligence and Legal Decision-Making: The Wide Open?' (2019) 36(5) Journal of
International Arbitration 21.
 insults. An AI system's recommendation or assessment might inadvertently compromise a
party's willingness to settle due to its inability to understand the significance of face-saving.
• Use of Written vs. Oral Evidence: AI's reliance on data means it might inherently favor written
evidence. In many Western legal traditions, written evidence is the cornerstone of the
arbitration process. However, in several Eastern and some non-Western cultures, oral evidence
holds primacy. While AI can process spoken words, it cannot effectively gauge the nuances of in-
person interactions, or the interpersonal dynamics present in direct testimonies or cross-
examinations. When AI aids in international arbitration involving parties from diverse traditions,
it's crucial to be aware of its limitations in fully appreciating the weight and importance of oral
evidence.
• Proverbs and Idioms: While AI has the potential to analyze vast amounts of data, understanding
the profound meanings behind proverbs and idioms remains a challenge. Misinterpretations can
compromise the integrity of the arbitration process.
• Cultural Norms and Taboos: Legal outcomes can be impacted if AI doesn't recognize what's
considered respectful in one culture versus another. The challenge lies in AI’s ability to
contextualize respect across cultures. While AI systems excel at rapid decision-making based on
data, they often lack understanding of norms and taboos. In certain cultures, consensus might
be sought through consultations with a broader group, potentially appearing to AI as inefficient
or redundant, without recognizing the cultural importance of such consultative processes.
• Emotional Nuances: In international arbitration, the stakes are high. An incorrect interpretation
of sarcasm or passive-aggressiveness by AI can lead to significant misunderstandings.

Legal and Ethical Implications:

• Limited Training Data: If AI operates without a holistic understanding, due to insufficient cultural
communication data, it raises questions about its fairness and credibility in arbitration.
• Absence of Empathy: Ethically, leaning on a tool devoid of empathy in situations demanding
emotional intelligence can be viewed as irresponsible. It can also pose legal challenges if one
party feels misunderstood due to AI's inability to recognize cultural emotions.
• Over-reliance on Stereotypes: Using AI that perpetuates stereotypes is not only ethically
problematic but can also lead to biased legal outcomes, affecting the credibility and legitimacy of
the arbitration process.
Accountability and Transparency Challenges in AI Decisions
Arbitration decisions must be based on clear reasoning and evidence, ensuring that parties
understand the basis for judgments and can challenge or appeal them if necessary. The black box nature
of many AI models runs counter to these principles, raising valid concerns about their appropriateness in
legal settings.

The complexity of AI systems, particularly those using machine learning algorithms, creates
obstacles in establishing accountability. In addition to the issues of 'black boxes', where the decision-
making processes are opaque and difficult to interpret, the complexity of some systems can make it
challenging to understand how the AI arrived at a particular decision. Additionally, since AI systems can
learn and adapt over time based on new data and experiences, unexpected decision-making outcomes
can occur that were not explicitly programmed by the developers.141 Under these circumstances,
establishing liability in AI-driven decisions is challenging. There can be many parties involved in the
development and application of AI, including software developers, data providers, operators, and users.
This diffusion of responsibility can make it challenging to determine who should be held accountable
when things go wrong.142 In many cases, the responsibility for an AI decision may be shared among
several parties. This distributed responsibility can further complicate the identification of liable parties
and decrease interest/trust in AI systems.

Moreover, many AI algorithms use high-dimensional vectors for their calculations, making it hard
for humans to understand the semantics behind the outputs. This inherent opaqueness of AI systems
may cause significant complications in the context of dispute resolution, where transparency,
interpretability, and understandability are crucial for the trust and acceptance of the decisions.

Many jurisdictions have privacy and data protection laws that address transparency in a broad
sense. For instance, the EU’s General Data Protection Regulation (GDPR) mandates transparency in
automated decision-making processes, including those powered by AI.143 Under Article 22 of GDPR, data
subjects have the right not to be subject to a decision based solely on automated processing, including

141
K Yeung, '“Hypernudge”: Big Data as a mode of regulation by design' (2017) 20(1) Information, Communication
& Society 118-136.
142
E P Rusakova and E Young, 'The Impact of Digital Technologies on Arbitration Courts' in Advances in Natural,
Human-Made, and Coupled Human-Natural Systems Research: Volume 2 (Springer International Publishing 2023)
445.
143
GDPR, art 5(1)(a).
profiling, which produces legal effects concerning them.144 Additionally, GDPR’s Recital 71 requires
meaningful information about the logic involved in the automated decision.145

In the United States, while there is no single, comprehensive federal law regulating the collection
and use of personal data, individual states such as California have enacted privacy laws with certain
transparency requirements.146

However, existing transparency laws are often found to be inadequate in regulating AI

technologies due to their limited scope and adaptability to fast-evolving AI technologies.

Data Privacy and Security Concerns in AI Systems

The integration of Artificial Intelligence (AI) in various domains has amplified the importance of
data privacy, especially in areas like arbitration where sensitive data is frequently processed. AI tools,
designed to enhance efficiency in arbitration, simultaneously bring forth challenges relating to data
protection and compliance.

For the purposes of contract analysis, AI tools are able to rapidly process thousands of contracts,
identify patterns and potential issues. Similarly, AI has been used in legal case analysis where it can
search and extract relevant information from vast databases of previous legal cases. However, the use of
AI in processing such sensitive information presents significant privacy implications.

The use and transfer of contract data and other pertinent data is inherently linked to privacy
rights of individuals and entities, necessitating rigorous data privacy regulations.147 These regulations are
not limited to digital data. Even when AI tools process information from paper filing systems or other
non-digital mediums, they must ensure that data privacy standards are upheld.

One significant challenge posed by AI integration in arbitration is the cross-border transfer of

personal and/or select organizational data.148 In the context of AI in arbitration, the software or tool's
provider might be viewed as the data controller, taking primary responsibility for adhering to data laws.
This includes ensuring lawful data processing, data minimization, and robust data security measures.
Meanwhile, those using AI tools to process data, potentially the arbitrators or legal counsel, might be

144
GDPR, art 22.
145
GDPR, recital 71.
146
California Consumer Privacy Act 2018.
147
G L Benton and S K Andersen, 'Technology Arbitration Revisited' (2020) 74 Dispute Resolution Journal 1.
148
K Noussia, Confidentiality in International Commercial Arbitration (Springer 2010) 22.
deemed as data processors with their distinct obligations. The landscape gets even murkier with the
concept of "joint controllers", where multiple entities collaboratively decide on data processing
methods. This collaboration demands clear delineation of duties and responsibilities to guarantee
compliance.

Most countries have data protection laws that grant data subjects specific rights, such as the
right to erasure or cessation of processing. As such, AI tools in arbitration must be equipped to handle
any grievances from data subjects, respecting and upholding their rights.

When multiple stakeholders – including legal teams, institutions, arbitrators, and AI tool
providers – handle personal data in the arbitration process, the demarcation of responsibilities becomes
intricate. The blending of AI tools into the arbitration framework underscores the intricacy of
maintaining data privacy in a scenario rife with multiple stakeholders and overlapping obligations.

Different countries have different laws and standards when it comes to data privacy. The
European Union, for instance, adheres to the GDPR, which imposes stringent requirements on how
entities process and transfer personal data.149

Historically, U.S. data privacy laws have been based on a "harms-prevention-based" approach,
focusing on preventing or mitigating specific harms in certain sectors. In contrast, GDPR adopts a "rights-
based" approach, where individuals effectively own their personal information and have the legal right
to control its use. Inspired by the GDPR, states like California, Colorado, Connecticut, Utah, and Virginia
are introducing new data privacy laws in 2023.150 These laws signify a move from sector-specific
regulations to a more comprehensive approach, encompassing various industries. The U.S. has
traditionally allowed businesses to collect personal information without explicit consent, regulating its
use to prevent harm in sectors like finance, health, education, and children's online activities. The new
U.S. state laws in 2023 reflect the influence of the GDPR's rights-based philosophy, introducing
comprehensive privacy protections across multiple sectors.151

Despite recent positive changes, continued disparities raise critical issues around AI systems and
the ability to maintain consistent standards of data privacy. The heterogeneity in international data

149
GDPR, art 5(1)(a).
150
F Bellamy, ‘U.S. Data Privacy Laws to Enter New Era in 2023’ (Reuters, 12 January 2023)
<https://www.reuters.com/legal/legalindustry/us-data-privacy-laws-enter-new-era-2023-2023-01-12/> accessed 14
August 2023.
151
ibid
privacy laws poses significant conflicts and compliance issues for AI applications which often transcend
national borders. Abiding by the regulatory frameworks of multiple jurisdictions can be challenging
which is further complicated by cross-border data transfer restrictions present in many privacy laws,
which can impede the global operation of AI systems.

Regulatory Challenges in AI Deployment

The rapid evolution and deployment of AI tools in various sectors have underscored the pressing
need for robust regulatory frameworks. Especially in the realm of arbitration, where AI tools capture,
analyze, and produce outputs from vast amounts of data, there's a critical need to ensure that these
processes are transparent, accountable, and respect the rights of all involved parties.152 Given that they
learn from this data, there's an inherent risk of these tools inadvertently revealing patterns or
information that parties intended to remain confidential. When AI tools process data that then gets
transferred across borders, there is potential exposure to varied data privacy regulations and
vulnerabilities.153 Parties must be assured that their rights under prevailing data privacy regulations, such
as the right to access, rectify, and delete their data, are maintained even when AI tools are in play.

Often, AI tools used in arbitration may be third-party products. The relationship and obligations
of these AI vendors concerning data confidentiality and security become a crucial element of the privacy
discourse. Under product liability laws, manufacturers could potentially be held liable for the harmful
outcomes produced by their AI systems.154 However, given the complexities and autonomous nature of
AI systems, it can be challenging to establish such liability.

In negligence law, the liability falls on those who failed to exercise reasonable care. However, in
the case of AI, it is often hard to determine what constitutes 'reasonable care'.155

In instances where AI is used to resolve disputes involving intellectual property, there can be
issues concerning copyright or patent infringement. Current laws may not sufficiently cover AI-created
works or inventions.

152
F Spyropoulos and E Androulaki, 'Aspects of Artificial Intelligence on E-Justice and Personal Data Limitations'
(2023) 26(3) Journal of Legal, Ethical and Regulatory Issues 1-8.
153
A Majeed and SO Hwang, 'When AI meets Information Privacy: The Adversarial Role of AI in Data Sharing
Scenario' (2023).
154
D C Vladeck, 'Machines without Principals: Liability Rules and Artificial Intelligence' (2014) 89 Washington Law
Review 1.
155
A Renda, 'Artificial Intelligence: Ethics, Governance and Policy Challenges' (CEPS Task Force Report 2019).
 As these complexities around AI in arbitration continue to mount, it becomes clear that there is a
need for a more comprehensive and evolved framework for the regulation of AI globally. The challenges
in establishing responsibility, protecting privacy, and ensuring proper use have caused governments
worldwide to establish regulations that balance innovation with privacy and safety.

While countries are at various stages of developing their approach to regulating AI there has
been legal and legislative action in several countries/regions.

In the European Union, the AI Act (AIA) was introduced in April 2021. This legislation proposes a
risk-based approach to AI use in both private and public sectors, categorizing risks into three levels:
unacceptable, high, and not explicitly banned. The AIA is still under review in the European
Parliament.156

Canada introduced the AI and Data Act (AIDA) in June 2022. This proposed legislation aims to
regulate the responsible design, development, and deployment of AI systems, specifically those that are
"high-impact" including systems related to civil and human rights and health. The Bill specifies those
tools that might lead to discrimination on prohibited grounds recognized by the Canadian Human Rights
Act. It is safe to assume, AI tools used in arbitration proceedings would be considered "High impact" and
would therefore need to identify, assess, and mitigate potential harms, ensure human oversight, provide
transparency, ensure fairness, and safeguard from misuse.157

The United States has not yet passed federal legislation governing AI. Instead, the Biden
Administration and the National Institute of Standards and Technology (NIST) have published broad AI
guidance. However, several US States have addressed the issue in the legislative session of 2023. At least
25 states, Puerto Rico, and the District of Columbia introduced AI-related bills and 14 states and Puerto
Rico either adopted resolutions or enacted legislation:158

156
‘EU AI Act: First Regulation on Artificial Intelligence: News: European Parliament’ (EU AI Act: first regulation on
artificial intelligence | News | European Parliament, 14 June 2023)
<https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-
artificial-intelligence> accessed 9 September 202
157
Bill C-27, 'Digital Charter Implementation Act, 2022: An Act to enact the Consumer Privacy Protection Act, the
Personal Information and Data Protection Tribunal Act and the Artificial Intelligence and Data Act', 1st Sess, 44th
Parl, 2022 (first reading 16 June 2022) <https://www.parl.ca/DocumentViewer/en/44-1/bill/C-27/first-reading>
accessed 11 September 2023.
158
K Zhu, ‘The State of State AI Laws: 2023’ (EPIC, 3 August 2023) <https://epic.org/the-state-of-state-ai-laws-
2023/#:~:text=Of%20the%20AI%2Drelated%20laws,blazes%20a%20path%20for%20A.I.> accessed 14 September
2023.
 Infographic: US State Legislations on AI
Regarding the use of AI Arbitrators, State laws in the U.S. provide varying interpretations. As
early as 1868, legal precedent described an arbitrator as “a private extraordinary judge chosen by the
parties in dispute, with the authority to settle the matter.”159 Although this definition predates modern
computer technology and algorithms, the term "private extraordinary judge" is sufficiently broad,
potentially accommodating advancements like AI technologies. In a similar vein, the Supreme Court of
Wisconsin embraced a definition of arbitrators as “one with the absolute power of deciding disputes
binding the disputants,” a scope that might include AI systems. However, it's noteworthy that a majority
of state courts, when discussing arbitrators, refer to them using the term “person,” suggesting the
recognition of an Automated Arbitration Intelligence (AAI) entity as an e-person.160

State-level arbitration rules and guidelines, however, seldom directly address the qualifications
or criteria for arbitrators. For instance, the rules set by FINRA characterize both public and private
arbitrators explicitly as “a person.” This lack of clarity poses questions about the acceptability and role of
AI arbitrators under current state laws.

China has actively addressed the challenges and risks associated with artificial intelligence (AI)
through the introduction of a series of regulations. The regulations have distinct focuses. The Algorithm
Recommendation Regulation centers on the application of algorithm recommendation technologies,
including generative and synthetic algorithms, in internet services. The Deep Synthesis Regulation

159
Gordon v United States (1868) 74 US 188, 194 (quoting Bouvier’s Law Dictionary, title “Arbitrator.”).
160
State ex rel. Cushion v City of Massillon (2011) Ohio App LEXIS 3922; Konz v Morgan Stanley Smith Barney (2018)
US Dist LEXIS 180069.
specifically addresses the use of deep synthesis technologies, a subset of generative AI, in internet
services. The Generative AI Regulation oversees the development and application of all generative AI
technologies for public services, excluding those not intended for public use. Meanwhile, the Draft
Ethical Review Measure is set to focus on the ethical review of AI technology R&D in China.

The overarching aim of these regulations is to manage the potential risks associated with AI-
generated content and to ensure national and social security in China. They impose specific duties on a
range of entities, from service providers and technical supporters to users and online platforms. For
instance, entities are required to file relevant algorithms with the Cyberspace Administration of China
(CAC) or its local counterparts within 10 working days of service provision. Moreover, before launching a
product or service, entities must conduct a security assessment in accordance with the Security
Assessment Regulation and submit a report to the local city-level cyberspace administration and public
security authority. Online app distribution platforms are also obligated to verify the completion of these
security assessments and filings before launching applications.161

Brazil is working on its first AI regulation law, with a draft that shares similarities with the EU's
draft AI Act. The draft focuses on three central pillars: guaranteeing the rights of people affected by the
system, classifying the level of risk, and predicting governance measures for companies that provide or
operate the AI system.162

Japan's AI strategies and regulations are closely tied to the major project "Society 5.0", which
aims to address social problems with innovation. The country has published the Social Principles of
Human-Centric AI and the AI Utilization Guidelines.163

In India, there is currently no specific regulatory framework for AI systems, but the government
has shown its intention to move forward with AI regulation through some working papers and it
Switzerland plans are underway to apply and adapt existing laws rather than creating a separate law for
the regulation of AI.164

161
M Sheehan, ‘China’s AI regulations and how they get made’ (Carnegie Endowment, 10 July 2023)
<https://carnegieendowment.org/2023/07/10/china-s-ai-regulations-and-how-they-get-made-pub-90117>
accessed 10 September 2023.
162
B Kohn and F-U Pieper, ‘AI Regulation around the World’ (Taylor Wessing, 9 May 2023)
<https://www.taylorwessing.com/en/interface/2023/ai---are-we-getting-the-balance-between-regulation-and-
innovation-right/ai-regulation-around-the-world> accessed 7 September 2023.
163
ibid
164
ibid
 While jurisdictions are indeed addressing the issue, the inherent challenges of AI governance in
arbitration, emphasizing the pivotal roles of rationality, controllability, and understanding the societal
implications.

Summary of Findings
The primary objectives of this research were to explore the current applications of AI in
international disputes, assess its potential to alleviate court burden, and identify the challenges that
come with it. The findings from this research could potentially guide legal professionals, policymakers,
and AI developers in leveraging AI for dispute resolution, while also being cognizant of the potential
challenges and ethical implications.

An essential observation from the findings is the importance of recognizing the supportive role
AI is already playing in international arbitration. It is crucial not to solely perceive AI in the context of
"robot arbitrators" but to understand its broader contributions in enhancing the arbitration process. This
perspective ensures a balanced view of AI's capabilities and limitations in the realm of international
arbitration.

The Finds show AI is a continuation of the integration of technology into international arbitration
practices which has surged due to the COVID-19 pandemic. This surge has led to a rise in virtual
arbitration proceedings, online dispute resolution platforms, and e-filing systems. These technological
inclusions have improved the efficiency and accessibility of arbitration processes. The shift towards
technology in arbitration is likely irreversible, influencing other areas like law firm operations and
university teachings. Additionally, technology has facilitated enhanced rules for evidence and disclosure,
with AI playing a pivotal role in the rigorous evidence and disclosure rules.

This continued integration has brought several key benefits to International Arbitration such as
enhancing efficiency and speed thereby ensuring swift resolution of disputes. AI, through technologies
like Machine Learning and Natural Language Processing, can analyze thousands of documents quickly,
outpacing human capabilities. AI also improves accuracy and consistency, removing human errors and
biases, fostering trust. Furthermore, AI can be cost-effective by minimizing human expenditure and can
augment accessibility and transparency in international arbitration.

While the findings recognize the current potential of AI in International Arbitration, they also
touch on the transformative possibilities through AGI (Artificial General Intelligence) - with its superior
cognitive abilities for strategizing, negotiation, or advising.
 Despite its advantages, AI's application in international arbitration has challenges. One of the
primary concerns is AI's current inability to fully replicate or substitute human intelligence, judgment,
and ethical considerations. Currently, AI lacks emotional intelligence, which is crucial for roles like
lawyers, arbitrators, and judges, especially in matters requiring the interpretation of the law or the
assessment of parties' intentions. Additionally, the Findings recognize the significant ethical and legal
issue of AI bias, often arising from flawed data inputs. AI bias can compromise the principles of fairness
and neutrality in international arbitration. Furthermore, the black box nature of many AI models
challenges the principles of accountability and transparency in AI-driven decisions. Arbitration decisions
must be clear and evidence-based, but the opaque nature of AI models can make this difficult.

5. Recommendations
Limiting AI Bias
Preventing bias in AI-powered arbitration system should involve a multi-pronged approach, with
measures related to data collection, algorithm design, cultural awareness, and system testing. Striking
the right balance between technology and human intuition is paramount to ensuring fairness and
cultural sensitivity in the arbitration process.

One of the keystones in ensuring fairness is the design of the algorithm. It should not merely
process data but be adept at recognizing and mitigating potential biases. This can be achieved through
techniques such as fairness-aware machine learning. Coupled with this, the AI systems need to undergo
rigorous system testing. Periodic audits and impact assessments become invaluable tools in identifying
and rectifying biases. Explainability techniques further augment this process by offering a clearer insight
into the AI's decision-making mechanism, enhancing transparency and fostering a sense of
accountability.

However, even the best algorithms can falter if the data they operate on is flawed. Thus,
meticulous attention must be paid to data collection. The data sets used to train AI should be
representative of all relevant groups, ensuring the absence of prejudiced assumptions or inadvertent
errors. Training AI on such diverse datasets not only ensures a holistic model representation but also
diminishes the chances of the AI making biased decisions.

Cultural sensitivity remains at the heart of international arbitration, given its diverse
stakeholders. AI systems must be adept at grasping the subtleties and intricacies of varying cultures to
deliver judgments that are not only fair but culturally cognizant. This requires AI algorithms to be trained
effectively to avoid misinterpretation of cultural nuances. It's also where human experts play a pivotal
role, providing oversight to AI decisions and ensuring they resonate with cultural sensitivity and
appropriateness. Building trust with stakeholders hinges significantly on this, as they are more inclined to
adopt an AI system they believe respects and comprehends their cultural context.

The vast data-processing prowess of AI, while impressive, can sometimes miss the subtleties that
human intuition, experience, and context-specific insights bring to the table. Hence, a delicate balance
between technological efficiency and human intuition becomes a necessity in the arbitration process.
Additionally, ethical oversight by humans ensures that AI decisions remain grounded, especially in
scenarios that weren't foreseen during its training.

Transparency is not just about making AI operations clear; it's also about ensuring accountability
for its actions. Established guidelines serve as a beacon for both developers and users, laying the
groundwork for transparent operations and a system of checks and balances. Reinforcing this approach is
the involvement of third-party audits. Regular assessments by independent experts can provide an
unbiased evaluation of the AI system, ensuring its adherence to cultural sensitivity and the absence of
biases.

Improving Accountability and Transparency

In international arbitration, the principles of fairness, accountability, and transparency are
sacrosanct. Arbitration decisions must be based on clear reasoning and evidence, ensuring that parties
understand the basis for judgments and can challenge or appeal if necessary. As mentioned earlier, the
black box nature of many AI models runs counter to these principles, raising valid concerns about their
appropriateness in legal settings.

However, the "Explainable AI" (methods and techniques that make the decisions and actions of
AI systems understandable to humans) provides a promising avenue to address the black box problem
inherent in many AI systems. The objective of XAI is to produce clear, coherent, and interpretable
explanations for each decision an AI system makes, allowing for greater trust and transparency. XAI
provides insights into how and why specific decisions were reached, ensuring that legal decisions made
or informed by AI are transparent, fair, and based on understandable criteria.

If an AI-driven decision is challenged, XAI allows for an audit of the decision-making process,
ensuring that the AI system can be held accountable. Additionally, by understanding the factors
influencing an AI's decision, legal practitioners can use this information to improve the quality and
accuracy of future decisions. Specific recommendation regarding improving accountability and
transparency include:

• Legal Domain Modeling: Before implementing AI tools, it's crucial to have a comprehensive
model of the legal domain in question. For instance, the paper focused on Article 6 of the
European Convention on Human Rights. Such domain-specific models ensure that the AI tools
are tailored to the unique requirements and nuances of the legal area they are meant to serve.
• Grounding in Legal Reasoning: AI tools should provide explanations that are deeply rooted in
legal reasoning. This ensures that the explanations are not only technically accurate but also
resonate with legal professionals, making them more trustworthy and usable.
• User-Centric Design: The design and implementation of AI tools should prioritize the needs and
preferences of the end users. This involves iterative testing and refinement based on user
feedback, ensuring that the tools are intuitive and effective in real-world legal settings.
• Transparency and Trustworthiness: AI tools should be transparent in their decision-making
processes. This transparency fosters trust, especially in critical domains like law where decisions
can have significant consequences.
• Continuous Evaluation: Regular evaluation exercises, involving real-world end users, are
essential. These evaluations not only gauge the effectiveness and usability of the tools but also
highlight areas for improvement.
• Addressing Limitations: While AI tools can be highly accurate, it's essential to recognize and
address their limitations. For instance, machine learning approaches might suffer from biases
inherent in the training data. Therefore, combining traditional knowledge representation
techniques with machine learning can offer a more balanced and robust solution.
• Emphasis on Explanation: In legal contexts, the "why" behind a decision is as crucial as the
decision itself. AI tools should, therefore, prioritize providing detailed explanations for their
conclusions, ensuring that users can understand and trust the provided insights.

Ensuring Data Privacy

AI Tools must incorporate both technical and organizational strategies to safeguard data, ensuring its
confidentiality and integrity. Moreover, in case of any data breaches, AI solutions should have
mechanisms to promptly address the issue, notify stakeholders, and implement remedial actions. Parties
involved in AI-powered arbitration must be meticulous in documenting any restrictions on transferring
personal data to third countries. They must also strategize ways to facilitate such transfers while
complying with data protection regulations.

• Specialized Expertise in Arbitration: Given the complexity of technology-related disputes, there's

a need for specialized expertise in arbitration. This ensures that the nuances of such disputes are
understood and addressed appropriately.
• Emphasis on Privacy: Considering the sensitive nature of data and the potential risks associated
with breaches, arbitration proceedings should prioritize privacy. This can be achieved by
including confidentiality provisions and ensuring that arbitrators routinely enter confidentiality
orders.
• Efficient Resolution: While arbitration is seen as a more efficient alternative to litigation, there's
room for improvement. The process can be further streamlined to ensure quicker resolutions,
which is particularly crucial for rapidly evolving technology sectors.
• Adaptability to New Technologies: As technology continues to evolve, arbitration processes
should be adaptable to encompass new technologies and the unique challenges they present.
This includes understanding the implications of technologies like AI, blockchain, and quantum
computing on arbitration.

Enhancing AI Governance and Regulation

Given the inadequacies of existing laws in addressing the novel issues presented by AI, there is a
need to revise existing legal frameworks and/or craft new regulations to manage the unprecedented
issues emerging from AI's proliferation. For instance, as AI finds its way into products and services,
traditional concepts of liability and negligence need revisiting. There's a compelling case to be made for
setting new benchmarks for 'reasonable care' specifically tailored for AI applications. Moreover, crafting
comprehensive regulations that highlight AI-specific concerns, such as transparency, accountability,
privacy, and fairness, has become indispensable.

One area that demands particular attention is the integration of AI in arbitration. Navigating AI
governance in this sphere requires a nuanced approach. Nations and regions must look towards
establishing dedicated governance mechanisms for AI in arbitration, ensuring not only effective law
enforcement but also preemptive system evaluations. Equally important is investing in specialized
training on AI ethics and governance for all stakeholders involved. This becomes even more crucial when
one considers the myriad of stakeholders that participate in AI-powered arbitration. Facilitating global
dialogues can act as a bridge, filling any informational gaps and promoting the exchange of best
practices.

Further diving into the realm of AI in arbitration, there are two discernible angles from which
regulations should be approached. First, the algorithms that power these AI systems must be subjected
to rigorous standards of certification and accreditation. Their credibility should be reflective of their
reputation within the market. Drawing inspiration from the world of arbitration, where parties have the
freedom to choose the governing law for their dispute, a similar liberty could be extended to allow
parties to select the algorithm that would oversee their case. Some nations even advocate for the
creation of a national register of algorithms to further instill trust and ensure equitable AI deployments.
Second, when integrating a machine arbitrator into a tribunal panel, it's essential to remember that the
machine's decisions remain under human scrutiny. With two human arbitrators on the panel, the
machine arbitrator's decisions are not taken at face value and can be either endorsed or overruled.
Interestingly, this dynamic can be flipped, with AI playing a pivotal role in highlighting and potentially
correcting biases in human arbitrators. By juxtaposing a machine's decision against a human's, the
impartiality and independence of the latter can be better gauged.

Lastly, a critical aspect that cannot be overlooked is the lawfulness of decisions influenced or
made by AI, especially in the context of arbitration. Any AI-derived decision or arbitral award must be in
strict compliance with all relevant laws of the jurisdiction where it is issued and where enforcement
might be sought. The onus falls on the arbitration community to guarantee that the AI systems they
deploy operate in sync with these binding laws. In essence, as we further delve into the AI era,
establishing sound governance and regulation becomes paramount. Addressing these challenges
systematically will enable us to leverage AI's capabilities both responsibly and effectively.

6. Conclusion
The transformative power of technology in the realm of law is undeniable, with Artificial
Intelligence (AI) standing at the forefront of this revolution. In the intricate world of international
arbitration, AI possesses the potential to alleviate numerous challenges that have long plagued this
sector, from lengthy proceedings to inconsistencies in arbitral awards. However, as with any nascent
technological development, the full impact of AI's integration remains to be charted, both in its
capacities and its challenges.
 The literature reveals that international arbitration has long been a robust and evolving platform
for settling international disputes. Its processes and procedures have a rich historical context yet are
constantly adapting to global events and technological advancements. Aspects like the increasing push
for transparency and challenges related to international arbitration processes have heightened the need
for innovation, and AI emerges as a viable candidate in this realm.

As gleaned from the findings, AI offers significant advantages in international arbitration,

including enhancing efficiency, accuracy, consistency, and cost-effectiveness. The allure of AI is further
bolstered by its promise to foster accessibility and clarity, as evidenced by advancements in large
language models and other specialized AI models.

However, AI in international arbitration is not without its challenges. Concerns over AI-induced
biases, recognition of cultural nuances, transparency, and data security have emerged as significant
impediments. Ethical, legal, and practical implications of AI's role, particularly in cross-border disputes,
necessitate cautious exploration. AI systems are only as good as the data they are trained on, which
raises pressing concerns about fairness and transparency in arbitral decisions.

Taking a retrospective lens on the objectives of this dissertation, the current applications of AI in
international disputes are multifaceted, offering both immense potential and accompanying challenges.
While AI's merits in alleviating court burden are evident, its full adoption requires a concerted effort to
address its inherent limitations and challenges.

AI's Key Features: Machine learning algorithms, Natural Language Processing, and
argumentation mining are just some of the AI facets contributing to international arbitration. These
features aid in legal research, prediction of case outcomes, and the facilitation of online/virtual dispute
resolution and negotiation.

Limitations of AI: For AI to be deemed trustworthy in the domain of international arbitration, it

must overcome limitations concerning biases, recognition of cultural nuances, and ensuring
transparency in its decision-making processes.

Ethical and Legal Challenges: The research underscores the need for a comprehensive legal and
ethical framework that guides the deployment of AI in international arbitration. Ensuring fairness,
transparency, accountability, data privacy, and security will be paramount.
 Real-world AI Implementation: The successful adoption of tools like the eBRAM Platform and
and the growing acceptance of AI case/contract management support tools such as EviSort, signal a
positive trajectory for AI's role in international arbitration.

To harness the full potential of AI in international arbitration, stakeholders must be proactive in

addressing its challenges. Recommendations include robust efforts in mitigating biases, establishing clear
guidelines and regulations for transparency and accountability, and fostering collaboration between AI
developers, legal practitioners, and policy-makers. This synergy will be crucial to ensuring that AI serves
as a tool of augmentation, not replacement, upholding the sanctity and trust inherent in international
arbitration processes.

Additionally, in light of the findings, it is essential to acknowledge the supportive role AI is

already playing in international arbitration and not merely envision AI in terms of "robot arbitrators".

In conclusion, while AI undeniably possesses the potential to revolutionize international

arbitration, its integration must be meticulous, discerning, and ethically guided. Future research should
focus on real-world applications and case studies of AI in international arbitration to glean more
actionable insights. The nexus of AI and international arbitration, though ripe with promise, demands
prudence and a shared commitment to upholding the sanctity and integrity of legal processes and
procedures.
7. References
Cases
Alassini v Telecom Italia SpA (Case C-317/08) [2010] ECR I-2213.

BSG Resources Limited, Vale S.A. v Republic of Guinea (ICSID, [2022]).

Cofftea Trading see Hanaro Shipping v Cofftea Trading

Eli Lilly and Company v The Government of Canada, UNCITRAL, ICSID Case No UNCT/14/2 (16 March 2017).

Grays Harbor County v Williamson (1981) 96 Wn 2d 147, 156.

Hanaro Shipping v Cofftea Trading [2015] EWHC 4293 (Comm) [16].

Konz v Morgan Stanley Smith Barney (2018) US Dist LEXIS 180069.

Legaspy v Financial Industry Regulatory Authority, Inc, N.D. Ill. 2020 WL 4696818 (2020).

Loomis v State (Wis 2016) 881 N.W.2d 749, 767.

Sky Power v Iraero (LCIA, [2022]), (HKCFI, [2023]).

State ex rel. Cushion v City of Massillon (2011) Ohio App LEXIS 3922.

Tumey v Ohio 273 U.S. 510 (1927).

Vattenfall AB, Vattenfall Europe AG, Vattenfall Europe Generation AG v Federal Republic of Germany, ICSID Case No
ARB/09/6 (11 March 2011).

Statutes, Legislation, Cases, and Legal Instruments

Arbitration Act of Brazil LEI 9.307; Art 10.

Arbitration Act of Ecuador LEY 14-417; Art 19.

Arbitration Act of Peru LEY 1017 Art 20.

Art 1023 Rv

Bill C-27, 'Digital Charter Implementation Act, 2022: An Act to enact the Consumer Privacy Protection Act, the
Personal Information and Data Protection Tribunal Act and the Artificial Intelligence and Data Act', 1st
Sess, 44th Parl, 2022.

California Consumer Privacy Act 2018.

Colombian Arbitration Act LEY 1563; Art. 7.

Convention on the Recognition and Enforcement of Foreign Arbitral Awards (adopted 10 June 1958, entered into
force 7 June 1959) 330 UNTS 38 (New York Convention) art 5.

European Parliament, ‘EU AI Act: First Regulation on Artificial Intelligence’ (14 June 2023) <EU AI Act: first
regulation on artificial intelligence | News | European Parliament>.

French Code of Civil Procedure, Art 1450

GDPR, art 22.

ICDR, ‘ICDR Arbitration Rules 2021’ (International Centre for Dispute Resolution 2021).

International Chamber of Commerce, ‘ICC Arbitration Rules 2021’ (2021).

International Chamber of Commerce, 'ICC Dispute Resolution Statistics' (2021).

International Chamber of Commerce, 'ICC Guidance Note on Possible Measures Aimed at Mitigating the Effects of
the COVID-19 Pandemic' (2020) 2020(2) ICC Dispute Resolution Bulletin 1.

International Chamber of Commerce, 'Leveraging Technology for Fair, Effective and Efficient International
Arbitration Proceedings' (18 February 2022).

Law Commission, Smart legal contracts (Law Com No 401, 2021).

London Court of International Arbitration, ‘LCIA Arbitration Rules 2020’ (2020).

London Court of International Arbitration, 'LCIA Facts and Figures - 2019'.

Netherlands Arbitration Act 1450

Singapore Government, 'International Arbitration (Amendment) Act' (2021)

Singapore International Arbitration Centre, 'SIAC Annual Report 2019'.

UK Government, 'Arbitration Act 1996 (Amendment)' (2022).

UNCITRAL, 'Convention on the Recognition and Enforcement of Foreign Arbitral Awards' (New York, 1958) art III.

UNCITRAL, 'Model Law on International Commercial Arbitration 1985'.

UNCITRAL, 'Status: Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York, 1958)'.

UNCITRAL, 'Transparency in Treaty-based Investor-State Arbitration' (2013).

UNESCO, 'Recommendation on the Ethics of Artificial Intelligence' (2022).

United Nations Convention on International Settlement Agreements Resulting from Mediation (signed 7 August
2019, entered into force [date of entry into force]) UN Doc V1900316.

United Nations, 'United Nations Convention on Transparency in Treaty-based Investor-State Arbitration' (2015).

US Government, 'Federal Arbitration Act (Amendment)' (2022).

Books and Reports

Aidid, A and Alarie, B, The Legal Singularity: How Artificial Intelligence Can Make Law Radically Better (University of
Toronto Press 2023).
Ashley, K, Artificial Intelligence and Legal Analytics: New Tools for Law Practice in the Digital Age (Cambridge
University Press 2017).

Born, G, International Commercial Arbitration (2nd edn, Kluwer Law International 2014).

Brekoulakis, S, 'On Arbitrability: Persisting Misconceptions and New Areas of Concern' in Mistelis, L & Brekoulakis, S
(eds), Arbitrability: International and Comparative Perspectives (Kluwer Law International 2009)

Copeland, B, 'Artificial Intelligence' in Zalta, E N (ed), The Stanford Encyclopedia of Philosophy (Winter 2019 Edition)
(2019).

Dolzer, R and Schreuer, C, Principles of International Investment Law (2nd edn, Oxford University Press 2012)

Landon, T and von der Weid, K, 'The Impact of COVID-19 on International Arbitration Procedure' in Landon, T and
von der Weid, K (eds), The Impact of Covid on International Disputes (2022) 84

Leith, P and Hoey, A, The Computerised Lawyer: a guide to the use of computers in the legal profession (Springer
Science & Business Media 2012)

Lew, J, Mistelis, L and Kröll, S, Comparative International Commercial Arbitration (Kluwer Law International 2003).

Moscati, M and others, Comparative Dispute Resolution (Edward Elgar 2020) 339

Noussia, K, Confidentiality in International Commercial Arbitration (Springer 2010) 22.

Osoba, O & Welser IV, W, 'An Intelligence in Our Image – The Risk of Bias and Errors in Artificial Intelligence' (Rand
2017) 13

Renda, A, 'Artificial Intelligence: Ethics, Governance and Policy Challenges' (CEPS Task Force Report 2019).

Rusakova, E P and Young, E, 'The Impact of Digital Technologies on Arbitration Courts' in Advances in Natural,
Human-Made, and Coupled Human-Natural Systems Research: Volume 2 (Springer International Publishing
2023) 445.

Susskind, R, Tomorrow's Lawyers: An Introduction to Your Future (Oxford University Press 2013) 45

Journal Articles
Bagherian-Marandi, N and Ravanshadnia, M, 'Two-layered fuzzy logic-based model for predicting court decisions in
construction contract disputes' (2021) 29 Artificial Intelligence and Law 453

Bathaee Y, 'The Artificial Intelligence Black Box and the Failure of Intent and Causation' (2017) 31 Harv JL & Tech
889.

Benton, G L and Andersen, S K, 'Technology Arbitration Revisited' (2020) 74 Dispute Resolution Journal 1.

Colorado, Li O, 'The Future of International Arbitration in the Age of Artificial Intelligence' (2023) 40(3) Journal of
International Arbitration 328.

D Carneiro and others, 'Online dispute resolution: an artificial intelligence perspective' (2014) 41 Artificial
Intelligence Review 211-240.

Commission, J, 'The increasing use of data analytics in international arbitration' (2020) New York Law Journal.
Dinnar, S and others, "Artificial intelligence and technology in teaching negotiation" (2021) 37(1) Negotiation
Journal 65

Dorsey, C, 'Hypothetical AI Arbitrators: A Deficiency in Empathy and Intuitive Decision-Making' (2021) 13(1)
Arbitration Law Review 12.

Fortese, F and Hemmi, L, 'Procedural Fairness and Efficiency in International Arbitration' (2015) 3(1) Groningen
Journal of International Law.

Karim, M S, 'Artificial Intelligence: An Undiscovered Future of Arbitration' (2019) 22(2) Int A.L.R. 47

Katz, D, Bommarito, M and Blackman, J, 'A general approach for predicting the behavior of the Supreme Court of
the United States' PloS one, 12(4), e0174698.

Majeed, A and Hwang, SO, 'When AI meets Information Privacy: The Adversarial Role of AI in Data Sharing Scenario'
(2023).

Marrow, P B, Karol, M and Kuyan, S, 'Artificial Intelligence and Arbitration: The Computer as an Arbitrator-Are We
There Yet?' (2020) 74 Dispute Resolution Journal 35

McCarthy, J, Minsky, M, Rochester, N and Shannon, C, 'A Proposal for the Dartmouth Summer Research Project on
Artificial Intelligence' AI Magazine, 27(4), 12 (1955).

McKendrick, J and Thurai, A, 'AI Isn’t Ready to Make Unsupervised Decisions' (2022) 15 Harvard Business Review.

Michaels, A C, 'Artificial Intelligence, Legal Change, and Separation of Powers' (2020) 88 U Cin L Rev 1083.

Nottage, L, 'Confidentiality and Transparency in International Arbitration: Asia-Pacific Tensions and Expectations'
(2020) 16(1) Asian International Arbitration Journal.

Paulsson, J, 'Arbitration in Three Dimensions' (2010) 60 International and Comparative Law Quarterly 291.

Prakken, H, Wyner, A, Bench-Capon, T and Atkinson, K, 'A formalization of argumentation schemes for legal case-
based reasoning in ASPIC+' (2015) 25 Journal of Logic and Computation 1141–1166.

Raghunathan, R, 'How do we break the language barrier in NLP' (LatentView Analytics, 2020)

Rayhan, S, 'Ethical Implications of Creating AGI: Impact on Human Society, Privacy, and Power Dynamics' (2023)
Artificial Intelligence Review

Ristovska, M, 'The Principle of Confidentiality In International Arbitration' (2021) 15(2) IBANESS Conference Series-
Plovdiv/Bulgaria 415-421

Roemmich, K, Schaub, F and Andalibi, N, 'Emotion AI at Work: Implications for Workplace Surveillance, Emotional
Labor, and Emotional Privacy' (2023) Proceedings of the 2023 CHI Conference on Human Factors in
Computing Systems 1-20.

Rosen, L, 'Law as Culture: An Invitation' (2011) 3 McGeorge School of Law Global Center for Business and
Development Annual Symposium.

Scherer, M, 'Artificial Intelligence and Legal Decision-Making: The Wide Open?' (2019) 36(5) Journal of International
Arbitration 21

Schmitz, A and Rule, C, 'OArb Enters the Age of Artificial Intelligence' (2023) 29 Dispute Resolution Magazine 36.

Spyropoulos, F and Androulaki, E, 'Aspects of Artificial Intelligence on E-Justice and Personal Data Limitations'
(2023) 26(3) Journal of Legal, Ethical and Regulatory Issues 1-8.
Srinivasan, R and González, BSM, ‘The Role of Empathy for Artificial Intelligence Accountability’ (2022) 9 Journal of
Responsible Technology 100021

Stipanowich, T J, 'Arbitration: The "New Litigation"' (2010) 1 University of Illinois Law Review 1

Surden, H, 'Machine Learning and Law' (2014) 89 Washington Law Review 87.

Vladeck, D C, 'Machines without Principals: Liability Rules and Artificial Intelligence' (2014) 89 Washington Law
Review 1.

Volokh, E, 'Chief Justice Robots' (2019) 68 Duke LJ 1135.

Yamane, N, 'Artificial Intelligence in the Legal Field and the Indispensable Human Element Legal Ethics Demands'
(2020) 33 Geo J Legal Ethics 877

Yeung, K, '“Hypernudge”: Big Data as a mode of regulation by design' (2017) 20(1) Information, Communication &
Society 118-136.

Websites and Other Online Sources

'AI-powered contract management software' (Evisort) <https://www.evisort.com/> accessed 27 June 2023.

'Arbitration Laws of the World' (International Arbitration Information, 8 July 2023) <https://www.international-
arbitration-attorney.com/arbitration-law-of-world/> accessed 29 July 2023

'Arbitrator Intelligence Platform' (Arbitrator Intelligence | Home) <https://arbitratorintelligence.vercel.app/>

accessed 11 August 2023

‘Civil Resolution Tribunal’ (British Columbia’s Civil Resolution Tribunal) <https://civilresolutionbc.ca/> accessed 1
September 2023

‘Contract Review Automation’ (Lawgeex, 3 November 2021) <https://www.lawgeex.com/cra/> accessed 11 August

2023.

‘ContractTracker’ (Donnelley Financial Solutions (DFIN)).

<https://www.dfinsolutions.com/products/ebrevia/contracttracker> accessed 22 August 2023

‘Online Dispute Resolution Powered by Modria’ (Tylertech) <https://www.tylertech.com/products/online-dispute-

resolution> accessed 17 August 2023

‘Practical Tools Platform’ (Kluwer Arbitration)

<https://www.wolterskluwer.com/en/solutions/kluwerarbitration/practical-tools> accessed 20 July 2023

Altlaw, 'What is sentiment analysis? using NLP in Ediscovery' (Altlaw eDiscovery, 2023)
<https://www.altlaw.co.uk/blog/what-is-sentiment-analysis-using-nlp-in-ediscovery> accessed 11 August
2023

'Arbitration Laws of the World' (International Arbitration Information, 8 July 2023) <https://www.international-
arbitration-attorney.com/arbitration-law-of-world/> accessed 29 July 2023

'Arbitrator Intelligence Platform' (Arbitrator Intelligence | Home) <https://arbitratorintelligence.vercel.app/>

accessed 11 August 2023

Bellamy, F, ‘U.S. Data Privacy Laws to Enter New Era in 2023’ (Reuters, 12 January 2023)
<https://www.reuters.com/legal/legalindustry/us-data-privacy-laws-enter-new-era-2023-2023-01-12/>
accessed 14 August 2023.
Brookhouse, O, 'Can Artificial Intelligence understand emotions?' (Telefonica, 17 April 2020)
<https://business.blogthinkbig.com/can-artificial-intelligence-understand-emotions/> accessed 27 June
2023

Burgess, M, 'UK Police are Using AI to Inform Custodial Decisions—But It Could be Discriminating Against the Poor'
The Wired (3 January 2018) <https://www.wired.co.uk/article/police-ai-uk-durham-hart-checkpoint-
algorithm> accessed 11 August 2023

Chalkidis, I, Androutsopoulos, I, and Aletras, N, 'Neural Legal Judgment Prediction in English' in Proceedings of the
57th Annual Meeting of the Association for Computational Linguistics (Association for Computational
Linguistics, Florence 2019) 4317, <https://www.aclweb.org/anthology/P19-1424> accessed 17 August
2023.

Corbett-Davies, S and others, ‘A Computer Program Used for Bail and Sentencing Decisions Was Labeled Biased
against Blacks. It’s Actually Not That Clear.’ (The Washington Post, 17 October 2016)
<https://www.washingtonpost.com/news/monkey-cage/wp/2016/10/17/can-an-algorithm-be-racist-our-
analysis-is-more-cautious-than-propublicas/> accessed 11 August 2023

Dastin, J, ‘Amazon Scraps Secret AI Recruiting Tool That Showed Bias against Women’ (Reuters, 10 October 2018)
<https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G> accessed
11 August 2023

eBRAM International Online Dispute Resolution Centre <https://www.ebram.org/services.html> accessed 22 July

2023.

'Ebravia app' <https://www.dfinsolutions.com/products/ebrevia> accessed 17 August 2023

Efstathiou, S and Apostol, M, 'Arbitration tech toolbox: Chatgpt - Arbitral assistant or fourth arbitrator?' (2023)
Kluwer Arbitration Blog <https://arbitrationblog.kluwerarbitration.com/2023/07/22/arbitration-tech-
toolbox-chatgpt-arbitral-assistant-or-fourth-arbitrator/> accessed 07 August 2023

International Bar Association, 'Rules on the Taking of Evidence in International Arbitration' (17 December 2020)
<https://www.ibanet.org/MediaHandler?id=def0807b-9fec-43ef-b624-f2cb2af7cf7b> accessed 7 July
2023.

Kira Systems, <https://kirasystems.com/> accessed 11 September 2023.

Kohn, B and Pieper, F-U, ‘AI Regulation around the World’ (Taylor Wessing, 9 May 2023)
<https://www.taylorwessing.com/en/interface/2023/ai---are-we-getting-the-balance-between-regulation-
and-innovation-right/ai-regulation-around-the-world> accessed 1 September 2023

Lex Machina, 'Legal Analytics' <https://lexmachina.com> accessed 6 August 2023

Litera, <https://www.litera.com; ThoughtRiver, <https://www.thoughtriver.com; Luminance,

<https://www.luminance.com all> accessed 6 August 2023

M de Westgaver, C, 'Canvassing views on AI in IA: The Rise of Machine Learning' (Kluwer Arbitration Blog, 12 July
2023) <https://arbitrationblog.kluwerarbitration.com/2023/07/12/canvassing-views-on-ai-in-ia-the-rise-
of-machine-learning/> accessed 10 September 2023

N, G and Hill, K, ‘Google’s Photo App Still Can’t Find Gorillas. and Neither Can Apple’s.’ (The New York Times, 22
May 2023) <https://www.nytimes.com/2023/05/22/technology/ai-photo-labels-google-apple.html>
accessed 11 August 2023.
Ozminkowski, R, 'Garbage in, Garbage Out' (Medium, 2021) <https://towardsdatascience.com/garbage-in-garbage-
out-721b5b299bc1> accessed 01 September 2023

PracticePanther, 'Law Practice Management Software' <https://www.practicepanther.com/> accessed 6 August

2023

Secure AI training, 'When on-premise beats the cloud' (Equus Compute Solutions, 2023)
<https://www.equuscs.com/secure-ai-training/> accessed 04 September 2023

Sheehan, M, ‘China’s AI regulations and how they get made’ (Carnegie Endowment, 10 July 2023)
<https://carnegieendowment.org/2023/07/10/china-s-ai-regulations-and-how-they-get-made-pub-
90117> accessed 10 September 2023

Silberg, J and Manyika, J, ‘Tackling Bias in Artificial Intelligence (and in Humans)’ (McKinsey & Company, 6 June
2019) <https://www.mckinsey.com/featured-insights/artificial-intelligence/tackling-bias-in-artificial-
intelligence-and-in-humans> accessed 11 September 2023.

SmartAdvocate, 'Legal Case Management Software' <https://www.smartadvocate.com/features/revolutionizing-

case-management-with-ai-technology/> accessed 6 August 2023

Son, H, 'This Software Does in Seconds What Took Lawyers 360,000 Hours' (The Independent, 28 February 2017)
<www.independent.co.uk/news/business/news/jp-morgan-software-lawyers-coin-contract-intelligence-
parsing-financial-deals-seconds-legal-working-a7603256.html> accessed 14 August 2023.

Thomson Reuters, 'HIGHQ Litigation Management'

<https://legal.thomsonreuters.com/en/products/highq/litigation-management> accessed 6 August 2023

UiPath Newsroom, 'UiPath Automates Facility for Conclusion of Arbitration Agreements to Support Business
Continuity for Companies, Arbitration Courts' (UiPath, 12 April 2020)
<https://www.uipath.com/newsroom/uipath-automates-facility-for-conclusion-of-arbitration-
agreements> accessed 4 August 2023

Wilkinson, A, Tsai, J and Curle, D, 'How natural language processing can improve legal search results' (Kira Systems,
202AD) <https://kirasystems.com/learn/how-natural-language-processing-improving-can-improve-legal-
search-results/> accessed 22 July 2023

Wong, D, 'Technology Assisted Review: How AI and Machine Learning Streamline the Review Process' (Veritone, 5
May 2023) <https://www.veritone.com/blog/technology-assisted-review/> accessed 1 August 2023

Wu, J, 'Empathy in Artificial Intelligence' Forbes (17 December 2019)

<https://www.forbes.com/sites/cognitiveworld/2019/12/17/empathy-in-artificial-
intelligence/?sh=7c5e0b7e6327> accessed 11 September 2023.

Zhu, K, ‘The State of State AI Laws: 2023’ (EPIC, 3 August 2023) <https://epic.org/the-state-of-state-ai-laws-
2023/#:~:text=Of%20the%20AI%2Drelated%20laws,blazes%20a%20path%20for%20A.I> accessed 14
September 2023.