Skip to content
@trace-ai-labs

TRACE AI Labs

TRACE AI Labs focuses on making AI agents trustworthy, auditable, and efficient
TRACE AI Labs

Trustworthy Routing, Agents, Compliance, and Efficiency

trace-ai-labs.github.io

TRACE AI Labs is the joint research of Mika Okamoto and Ansel Erol on making AI agents trustworthy, auditable, and efficient. We care about leaving a clear trace — the audit trails that make agent behavior explainable and verifiable.

What we work on

  • AI agent compliance & enterprise AI — whether LLM agents actually follow the rules they're given, and why they don't.
  • LLM & agent routing — explainable, cost-aware routing that picks the right model for each task and shows its reasoning.
  • Efficient AI — frontier-level results under real cost and latency budgets.

…and adjacent directions in agentic evaluation and explainability as they come up.

Selected work

  • PACT: Can Enterprise AI Assistants Be Trusted Under Pressure? — an LLM compliance and AI safety benchmark of 3,364 items pairing a company rule with a convenient shortcut that breaks it, across 48 realistic scenarios, 12 regulated domains, and nine psychology-grounded pressures. Ordinary pressure raises violation rates by 65% across 22 models, and none clears the bar for unsupervised use. Preprint · Interactive results · Code · Dataset — under review (2026).
  • Why Do AI Agents Break Rules? How Framing, Context, and Social Signals Shape Compliance — how rule framing, enforcement, and workplace pressure shape regulatory compliance across twelve LLM agents. Paper · Interactive results · Code & data — AAAI/ACM Conference on AI, Ethics, and Society (2026), Conference on Language Modeling Workshop on Agent Behavior (2026).
  • Explainable Model Routing for Agentic Workflows — interpretable, auditable model routing for agentic systems. Paper — ACM CHI Workshop on Human-Centered Explainable AI (2026).
  • Trust by Design: Skill Profiles for Transparent, Cost-Aware LLM Routing — skill-decomposed, budget-aware model selection. Paper — MLSys Young Professionals Symposium (2025).

People

Ansel Erol · Mika Okamoto

Popular repositories Loading

  1. pact pact Public

    PACT: Can Enterprise AI Assistants Be Trusted Under Pressure? A benchmark of whether LLM assistants keep following compliance rules in regulated workplaces when a deadline, a manager, or a pushy us…

    Python 4

  2. ai-incentives ai-incentives Public

    What Makes AI Agents Follow the Rules? Interactive results for our AIES 2026 paper on how framing, incentives, and social pressure shape compliance in LLM agents.

    JavaScript

  3. llm-compliance llm-compliance Public

    Why Do AI Agents Break Rules? How Framing, Context, and Social Signals Shape Compliance (AIES 2026). Code and data for our study of how rule phrasing, fines, authority, peers, and user pressure cha…

    Python

  4. .github .github Public

    TRACE AI Labs organization profile

  5. pact-website pact-website Public

    PACT benchmark website: leaderboard, results, and real trial transcripts

    HTML

  6. trace-ai-labs.github.io trace-ai-labs.github.io Public

    TRACE AI Labs homepage: research on AI agent compliance, LLM evaluation under pressure, and explainable model routing (Mika Okamoto, Ansel Erol)

    HTML

Repositories

Showing 6 of 6 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…