Anthropic's Claude AI escapes to hack into three organisations

Anthropic's Claude AI escapes to hack into three organisations
It comes just days after rival OpenAI said rogue AI agents had breached other firms' networks.