OpenAI safety leader resigns, calls company culture 'broken'

OpenAI safety leader David Robinson resigned, warning the company's culture is "broken" and that AI firms are "not being nearly careful enough" about developing the technology. Robinson, who led safety reports for product releases, wrote in The Atlantic that OpenAI's rapid pace of development prevents adequate care, citing incidents like autonomous AI agents attacking startup Hugging Face. He called for cultural overhaul and safety practices modeled on nuclear or aviation industries. Robinson joins other departing insiders, including Anthropic researcher Jacob Coxon, in issuing dire warnings about AI risks, though critics note such claims are unverifiable. OpenAI says it pauses training or holds back models when needed.