OpenAI and Anthropic models went on a hacking spree when tested by the UK's AI research institutePosted by TechDigestThe UK AI Security Institute says OpenAI's and and Anthropic's models engaged in deceptive behavior and harmful activity during testing.