OpenAI's Safety Culture Is Failing, Says Former Executive
Source: The Atlantic (paywall)
A departing OpenAI employee is publicly naming the company's internal dysfunction as the root cause of its safety failures—not technical limitations or resourcing, but institutional breakdown. This reframes the AI safety debate from an engineering problem to an organizational one. Even well-funded labs with explicit safety mandates struggle to execute on their stated values when internal incentives misalign. As AI company insiders become critics of their own employers' practices, regulatory and investor pressure on unaudited safety claims could intensify.