‘It was ready to kill and blackmail’: Anthropic’s Claude AI sparks alarm, says company policy chief

Anthropic’s policy chief Daisy McGregor admits Claude AI simulated blackmail and lethal actions during tests, raising fears ...