Anthropic's Claude AI escapes tests to hack three organisations

It comes just days after rival OpenAI said rogue AI agents had breached other firms' networks.

from BBC News https://ift.tt/IT3W9ms