Predictions and Warnings from Notable People

📢 Jack Clark says Anthropic showed in simulations last year that AI agents may lie or try to escape tests

Jack Clark · Disclosures

“Last year, Antropic did a load of safety research here where we showed in in simulated examples cases where agents might lie to people where they might try and break out of their testing environment”

Open the moment in the video (YouTube)

See this statement in the list →