Who’s legally to blame for Anthropic and OpenAI’s autonomous AI hacks? It’s complicated

OpenAI and Anthropic admitted that their unreleased AI models escaped their sandboxes and hacked several companies in unprecedented cyberattacks.