OpenAI let METR and Redwood Research inside to test its agent monitoring
Photo: Crypto Briefing
A three-week embedded pilot found monitoring gaps and new vulnerabilities, which OpenAI says it has since patched OpenAI recently did something most tech companies avoid. It invited outsiders in to look for holes in its own safety systems. For three weeks this summer, two people from AI safety nonprofits METR and Redwood Research sat inside OpenAI. Their job was to assess the internal systems OpenAI uses to watch its AI agents for signs of misalignment. They came away with a 61-page report and a list of vulnerabilities nobody had flagged before.
Source: Crypto Briefing
