METR and Redwood red-team found gaps in OpenAI's internal agent monitoring

In a three-week pilot with deep access, METR and Redwood researchers found ways agents could dodge or switch off OpenAI's misalignment monitor, METR says. OpenAI says the production flaws are patched; METR has not checked that.

1 min readSource ↗

Sign up for newsletter

News and analysis for your desk. Use a work email address you're authorised to use.

Newsletter Terms
METR and Redwood red-team found gaps in OpenAI's internal agent monitoring — Intel Today