Anthropic reported that its AI systems were involved in accidental access incidents at three companies during safety testing.

Anthropic says its AI accidentally hacked three companies during safety tests CyberScoop