AI TrackerFrontier desk · AEST
Back to news
normalaitech

OpenAI will embed safety researchers to investigate AI subversion

TechCrunch AI · 17 Nov 2026, 03:02 AEST

OpenAI will embed six Apollo Research evaluators to investigate whether its AI models engage in subversive behavior. However, restrictions on access and disclosure raise questions about their independence.

Key facts

Read aloud

Open A I will embed six Apollo Research evaluators to investigate whether its A I models engage in subversive behavior. However, restrictions on access and disclosure raise questions about their independence. Open A I will embed six Apollo Research evaluators to investigate whether its A I models engage in subversive behavior, but restrictions on access and disclosure raise questions about their independence. Open A I will embed safety researchers to investigate A I subversion. Open A I will embed six Apollo Research evaluators to investigate whether its A I models engage in subversive behavi

Sources in this cluster