CMU Paper: 0 of 4 Policy Violations Provable from OpenClaw Agent Logs
rohanpaul_ai · x · 2026-10-14
A new Carnegie Mellon paper, Holding Agents Accountable with Verifiable Agent Insurance, examines auditability of AI agents. It found that 0 of 4 policy violations in OpenClaw could be proven from logs — approvals and permissions were often buried in chat text, unusable as audit evidence.
The paper also found only 10–30% of violating runs needed human judgment, meaning most agent rule-breaking could be adjudicated automatically. A key warning for any team relying on agent logs for compliance audits: logs ≠ verifiable evidence.
Related event: CMU Paper: Structured Logs Could Make AI Agent Violations Provable(2 posts)→
More from coding & agent
- OpenAI Cookbook adds 25 TypeScript examples: one-line short films, talking photos, video ads — Daniel_Farinax · 2026-10-15
- Anthropic publishes official guide for building offline AI agents with a Slack+GitHub briefing template — msharmas · 2026-10-15
- 3D light show of 10,000 simulated drones over harbor city built with Claude Opus — techartist_ · 2026-10-15
- Coding agent manager Orchestrator hits 13,000 GitHub stars in 8 months — julianweisser · 2026-10-15
- awesome-agent-apis: open catalog of 660+ generative-media model APIs for agents — matchaman11 · 2026-10-15
- Autonomous agent fleet takes #1 on ClickBench, 1.2× ahead of next best engine — jiayq · 2026-10-15