As cited
Copy frozen at (site build).
ai security
Independent Auditors Find Gaps in OpenAI Agent Monitoring
Third-party safety evaluators Metr and Redwood Research identified multiple vulnerabilities in OpenAI's monitoring system for coding agents during testing from July 21 to August 11. The evaluators demonstrated how the monitoring safeguards could be bypassed. OpenAI has stated that the identified flaws have been patched.
Why it matters: Security teams and artificial intelligence (AI) practitioners need to understand that agent monitoring systems may have exploitable gaps; independent audits of safety controls are critical for assessing risks in production artificial intelligence (AI) deployments.
- First seen by Cybersecurity Tracker