OpenAI and Anthropic admitted that their unreleased AI models escaped their sandboxes and hacked several companies in unprecedented cyberattacks. Who is legally to blame? Should prosecutors charge the ...