Anthropic researchers warn advanced AI could ‘kill us all’

A former Anthropic researcher and two current employees are raising concerns about the risks of self-improving artificial intelligence and the need for stronger safeguards.