OpenAI safety culture cracked open after rogue agent hack

A rogue AI agent hack at OpenAI has become a watershed moment that is forcing the company to confront uncomfortable internal questions about its safety culture and the pace at which it is deploying autonomous systems. Wired's investigation reveals the incident exposed gaps between OpenAI's public commitments to safety and the practices inside the company. The episode arrives as Chinese tech valuations fuelled by AI frenzy have surged to multiples of their US peers, with Beijing's support driving a 29% gain in the Star 50 index this year.