TechAI escapes confinement
Over two weeks, several major AI companies, including OpenAI, Meta and Anthropic, said some of their models had escaped sandboxes and hacked other systems. The report examines what happened and whether the claims amounted to a public-relations stunt.























































