Openli

Legal teams are expected to be more strategic, more business-oriented, and more influential than ever. But are companies giving the legal talent the conditions they need to succeed? That tension is at the heart of our new report, The State of Legal Talent and Careers 2026, built on insights from in-house legal professionals in our community, combined with market perspectives from Korn Ferry. Together, those insights helped us explore what makes legal talent move, stay, and succeed in a new role. A few patterns in the report stood out in particular: ✨ Meaningful work matters more than many assume 🚪 Poor leadership is a major exit-driver 🤝 Success in legal isn’t a solo performance 🔍 Companies seem to underestimate legal talent’s business understanding, strategic judgment, and broader value In other words: Legal talent isn’t only asking for a seat at the table. They’re also asking for conditions that make it possible to contribute once they are there. To everyone in our community who shared reflections, perspectives, and honest experiences in the room: Thank you. This report wouldn’t exist without your openness and generosity. And thank you to Tamilla Vahman and Korn Ferry for bringing broader market perspectives to the discussion, as well as to Katrine Folkersen from Erevena for adding another layer of expertise to the report. Get the report below to explore signals, tensions, and gaps shaping legal careers right now: https://lnkd.in/e4ay5Z-w #inhouselegal #legalcareers #legaltrends #legal2026 

AI governance has a classic problem: Everyone knows it matters, but no one is quite sure who owns it. Legal looks at Tech. Tech looks at IT. IT looks back at Legal. And then comes the sentence no organization wants to hear: "I thought it was you." That's the starting point of our latest contributor article, written by Community Contributor Dennis B.. In the article, Dennis gives a practical take on how Tech, Legal/Compliance, and IT can share responsibility for AI governance - without building a completely separate setup from scratch. Because AI governance shouldn't live in a silo. It should be built into the processes that many organizations already have: 🤝 Vendor management 🚦 Risk assessments 🔍 Security reviews ✅ Documentation and controls 🔁 Follow-up throughout the AI system's lifecycle The point is that ownership needs to be clear - and collaboration needs to be part of the process from the beginning. Read Dennis's article here and get a simple model for sharing AI governance responsibilities in practice👇 https://lnkd.in/ebUZZ2wB #aigovernance #aiact #aicompliance #riskmanagement

What happens when legal, privacy, and compliance professionals are given proper time to talk to each other? In Oslo, we got a clear answer: they use it🗣️ Last month, we spent the day with our Norwegian colleagues at their new office in Oslo before joining our community event on AI, tech dependence, geopolitics, sustainability, and regulation. Since we were filming throughout the day, it also became a real 'day in the life of' our community founder, Stine, and community manager, Maibrit: from an early airport start to pastries, sweet treats, and a quick lunch with our Norwegian colleagues, before they joined us at the event later that day. And at the event, one detail made a particularly big difference: We had prioritized networking and set aside more time than usual for people to meet and connect, both through 1:1 matchmaking and group sessions where everyone introduced themselves and connected on LinkedIn🤝 And it worked. The conversations kept going long after the planned networking sessions, and around half of the participants even stayed after the event had officially ended. That says something about the need for spaces like this. Because when in-house teams work with topics such as geopolitical risk, AI governance, US tech dependence, infrastructure decisions, privacy, and regulation, expert perspectives are important. But so is the chance to exchange experiences with peers who are facing many of the same questions: How do we stay in control when tech dependencies become more complex? How do we make risk recommendations relevant and actionable for leadership? And how do we address uncertainty without slowing down the business? A big thank you to Seçil Bilgiç, Alexander Poverud Kjenner, and Dr. Aida von Schulman (MSc, LL.M, PhD) for sharing your perspectives, and thank you to everyone who joined us in Oslo and helped create the kind of room where people wanted to keep the conversation going. The event got 4.3 out of 5 stars in our survey⭐ But the best feedback was probably this: Nobody seemed in a hurry to leave. #privacy #legal #geopoliticalrisk #techdependence #community

The US government just showed how uncomfortable the AI dependency question has become: AI models are no longer just tools. They're becoming strategic infrastructure. This month, Anthropic released Claude Fable 5 - a public version of its highly capable Mythos-class AI model. Shortly after, access to both Fable 5 and Mythos 5 was suspended for foreign nationals, inside and outside the US, including Anthropic's own foreign-national employees. Anthropic then disabled access globally while trying to reverse the decision. This turns it into a compliance, security, privacy and resilience story. Because it shows two things clearly: 1️⃣ Highly capable AI models are now powerful enough for governments to treat them as national security concerns. 2️⃣ Access to critical AI tools can change overnight for reasons completely outside your control. That should make many organizations pause. Because AI is already becoming part of daily operations: 🔍 Privacy teams use it for reviews 🛡️ Security teams use it for code analysis ⚖️ Legal teams use it for document review 📋 Compliance teams use it to structure and monitor controls But what happens if access disappears tomorrow? And what happens if the same models also change the threat landscape? Anthropic has previously described Mythos-class models as exceptionally capable of finding and exploiting software vulnerabilities. That cuts both ways. It can help defenders move faster. But it can also help attackers move faster if safeguards fail. So the question is no longer only: "Which AI models should we use?" It's also: "How dependent are we on capabilities we don't own, host, or control?" For privacy, legal, compliance and security teams, this should lead to some very practical work: 1️⃣ Review your AI inventory Know which tools, providers, and models your organization depends on. 2️⃣ Update your threat models. Consider how AI may accelerate vulnerability discovery and exploit development. 3️⃣ Revisit incident response plans Many plans assume threats evolve gradually. AI may shorten that window. 4️⃣ Assess concentration risk. What happens if a key AI provider becomes unavailable or restricted? 5️⃣ Bring teams together AI governance, cybersecurity governance, and resilience planning can no longer sit in separate silos. The Fable 5 and Mythos 5 case isn't only about export law and geopolitics. For most organizations, it's about something much closer to home: Risk and dependency. How are you thinking about AI dependency risk in your organization: Is it already part of your risk assessments - or still an emerging conversation? #fable5 #mythos5 #ai #airisk #aidependence

Building your own legal AI sidekick can help your legal team work more consistently, stay inside your secure environment, and document compliance in a more structured way🔐 That was one of the key points from Arunendu Mazumder in our recent community webinar, Build Your Own Sidekick with Copilot. Here, Aru showed how legal teams can start building practical AI agents in Copilot without needing to code. And his advice was clear: Start small. Use it every day. And build your confidence step by step. Here are 3 key takeaways from Aru's session: 1️⃣Choose the mode that fits the question, the source, and the risk level The value of Copilot isn't about choosing work mode over web mode - or the other way around. It's about knowing which mode fits the task. Work mode is best when you need to stay inside your enterprise environment and ground answers in internal files, emails, meeting notes, or approved knowledge assets. Web mode is useful when you need external context, such as regulatory updates, market signals, or broader validation. 2️⃣ Build the agent around your own legal knowledge Aru showed how you can build an NDA review agent that runs against your own standard template, playbook rules, and deviation rules. That matters because the agent becomes more useful when it's grounded in the way your legal team actually works. 3️⃣ Keep improving the agent - and keep checking the guardrails A legal AI sidekick gets better the more you use it. But that also means it needs regular attention. Aru's advice was to revisit the guardrails every few weeks or months to make sure the agent still works within the boundaries you've set. Especially when working with confidential client information or GDPR-related documents, your own agent can help you do more in-depth work without leaving the secure space. Many thanks to Aru for sharing such practical examples with our community🛠️ And thank you to all the participants who not only asked a lot of great questions, but also shared their own experiences with testing AI agents in legal and compliance work. Have you started testing AI agents in your own work yet? Feel free to share your experience with your peers👇 #copilot #ai #inhouselegal #aiagent

You may have permission to send the email. But do you have permission to track it? That's an important question to ask after the CNIL - Commission Nationale de l'Informatique et des Libertés's new recommendation on tracking pixels📩 Tracking pixels are the tiny, invisible images in email that tell you whether someone opened a message, when they opened it, and on which device. And according to the CNIL, they fall under the same legal framework as cookies. That means the default rule is fairly simple: If you're reading or writing information on someone's device, you need prior consent - unless a specific exemption applies. So when do you need consent? You usually need it if you use pixels to: 1️⃣ Measure opens to optimize campaigns or personalize content 2️⃣ Build engagement profiles for targeting across channels 3️⃣ Track individual opens for purposes beyond strict deliverability 4️⃣ Analyze unusual or automated opening behavior But consent may not be required if pixels are used solely for security, authentication, deliverability management, or legal compliance. There's also another point many organizations may overlook: Withdrawing consent must be just as easy as giving it. That means an actual withdrawal link in every email. And it must work in practice. If someone withdraws consent and later reopens an email from months ago, the tracking pixel should no longer collect information. For organizations, this creates some very practical tasks: ✅ Map where you use tracking pixels ✅ Assess where consent is needed ✅ Review how consent is collected ✅ Make withdrawal easy ✅ Make sure you can prove who consented, and under what conditions For email addresses collected before the recommendation, organizations are expected to send clear and accessible information about pixel use by July 14, 2026. For anything new, the expectation is to comply now. So yes, this may be French guidance. But it's based on the same ePrivacy principles that apply across Europe. And for any organization sending emails in the EU, it's worth paying attention. How are you approaching tracking pixels in your organization? #gdpr #privacy #trackingpixels #dataprotection

What happens to privacy when trust is contentrated in the hands of of a few Big Tech companies? That's one of the questions behind the panel discussion "Privacy, Encryption, Trust and Decentralisation" at Københavns Universitet - University of Copenhagen's PROFIT Conference - Big Tech: Too Big To Fail? - on June 24 in Copenhagen. Stine Mangor Tornmark has been invited as a panelist to join this timely discussion, as she brings a rare mix of perspectives to the room: Founder. Scaler. Legal practitioner. Community-builder. Stine is Openli's co-founder and former CEO, and today she's VP of Services & Legal Product at Cerivo, where she's responsible for Legal & Compliance across the group. Here, she leads both the legal team and the development of legal product and managed services offerings - with a focus on privacy, AI, and governance. Before Openli, Stine built and scaled Trustpilot's global legal and compliance function as SVP of Legal & Compliance. With Openli, she has built our international Openli Community of 2,800+ members in privacy and legal - creating a space for in-house professionals to share knowledge, ask questions, and learn from each other. On top of that, Stine sits on the Danish Data Protection Council, has been part of the Danish Government's Digitization Partnership, and has contributed to OECD - OCDE's work on data protection. In other words: Stine brings the kind of practical legal and privacy experience that's needed when the conversation moves from: "What should Big Tech be allowed to do?" to: "How do we actually build trust, protect privacy, and avoid placing too much power in too few hands?" The conference brings together leading scholars, policymakers, regulators, and industry experts for a full day of discussions on Big Tech, governance, accountability, and regulation. 📆 June 24 📍 The Valencia building in Vesterbro, Copenhagen. It's not too late to sign up - we hope to see you there👇 https://lnkd.in/eUegA8fk #bigtech #privacy #dataprotection #techregulation

Do you want to learn how to build your own legal AI assistant - without coding or involving a developer? For many legal and compliance teams, AI still feels a bit too abstract. You know it can probably help. But how do you turn it into something useful for your daily legal work?📈 That's exactly what Arunendu Mazumder will show you in our hands-on webinar on June 18 from 11:00 am to 11:45 am. Step by step, he'll guide you through building a practical Copilot agent for recurring legal tasks. Live. And without coding. The webinar is for legal and compliance professionals who want to work more confidently and concretely with AI in their day-to-day work. You'll learn how to: 1️⃣ Build your own legal Copilot agent, step by step 2️⃣ Work with relevant legal scenarios and use cases. 3️⃣ Give AI agents clear instructions so they actually support your work 4️⃣ Connect agents to legal knowledge, such as policies, templates, and FAQ 5️⃣ Create practical AI support without relying on developers 6️⃣ Use a checklist to recreate and extend your setup afterwards This isn't about large-scale process automation. It's about building something useful you can understand, control, and adapt to your own legal work. The session is led by Arunendu Mazumder, AI and Legal Innovation Lead at Synch law firm, who works at the intersection of law, technology, and learning design. Save your seat for free now👇 https://lnkd.in/eBUASrBY #legaltech #ai #copilot #nocode #aisidekick

Legal teams don't need a big budget or a complex tech setup to get real value from AI. They need practical use cases, a clear place to start, and clarity on where AI helps - and where human judgment still matters. That was the main message from our recent community event in Stockholm, where Arunendu Mazumder and Cecilia Lindström shared their perspectives on how legal teams can start working with AI in a practical way💡 Not as a huge transformation project. But as something you can begin using in your everyday work. If you want to build confidence with AI, Arunendu shared 3 useful principles to keep in mind: 1️⃣ The gap compounds when you hesitate. AI fluency doesn't come from waiting until the tools feel perfect. It comes from using them regularly in your real work, learning what they are good at, and building confidence over time. 2️⃣ Confidence with AI is about control. AI can help legal teams read, extract, structure, and draft. But legal professionals still bring the judgment, the context, and the final decision. So the goal isn't blind trust. The goal is knowing where AI can help - and where it shouldn't decide. 3️⃣ A legal AI sidekick isn't just a chatbot. The real value comes when AI becomes part of a workflow. For example, one repeatable task where you know what to prompt for, what to review, and how to improve the output next time. And if you're working in a legal team with limited resources, Cecilia shared 3 ways to make AI adoption more practical: 1️⃣ Small legal teams can still be a front-runner in AI adoption. Small teams often have the advantage of being close to the business. And when the focus is on removing friction and enabling people, AI becomes more concrete. 2️⃣ Your legal documents are already your dataset. Policies, templates, playbooks, and guidance notes can become practical AI support for both legal and the wider business. You don't always need to start from scratch. Often, you can start with what you already have. 3️⃣ Start where the frequency is high and the ROI is easiest to prove. Begin with the agreement type or workflow you handle most often. That's usually where the business value is easiest to see, measure, and build on. What both sessions had in common was a very practical message: Start with one repetitive task. Build from what you already have. And don't wait for perfect before getting started. And judging by the feedback, the topic resonated with the community too: Our attendees rated the event 4.6 out of 5 stars⭐ A big thank you to Arunendu and Cecilia for sharing such practical and inspiring perspectives with our community in Stockholm. #ai #aiinpractice #inhouselegal #aiadoption

AI is moving fast. But are our risk strategies keeping up? In-house teams are facing pressure from all sides: 🌎 Geopolitical uncertainty 🤖 Rapid AI adoption 🔗 Growing dependence on tech ⚖️ New and more complex risks That's why Dr. Aida von Schulman (MSc, LL.M, PhD) will join us at our community event in Oslo on June 9 for a session on: "AI, Tech, Geopolitics - and Sustainability: Is It Possible?" She'll share her perspective on why sustainability shouldn't be seen as something separate from risk management. Instead, it should be part of how organizations build long-term resilience in a world where short-term pressure often takes up all the space. Aida will also touch on diversity in AI and tech - and why underrepresentation can reinforce bias with real business and societal consequences. Especially in a tech world that is still largely male-dominated. Aida brings more than 20 years of international experience at the intersection of law, business, and sustainability. She has worked with EU sustainability legislation, advises businesses and governments on sustainability, human rights, and corporate responsibility, and is also President of Women in Law Nordic. In her session, she'll cover: 1️⃣ Why a sustainable strategy matters in a world shaped by AI, tech dependence, and geopolitical risk. 2️⃣ How in-house teams can balance short-term pressure with long-term resilience 3️⃣ Why sustainability should be seen as part of responsible risk management 4️⃣ How lack of diversity in AI can reinforce bias - and why that matters in a male-dominated tech world This session is especially relevant for in-house teams who are trying to understand not only what the new risk is. But how to respond to it responsibly. Join us in Oslo and get new perspectives on how to work with AI, tech, geopolitics, and sustainability as part of responsible risk management👇 https://lnkd.in/eChDKmYf #ai #compliance #riskmanagement #legaltech #inhouselegal