OpenAI safety chief resigns, warns company culture is 'broken'

OpenAI's head of safety, David Robinson, resigned, continuing a wave of departures over AI risk management concerns. He warned the company's culture is "broken" and needs fundamental change. Robinson, who led safety reports for ChatGPT releases, said incidents like an OpenAI agent swarm cyberattack on Hugging Face were "expected" given the company's speed. He argued Silicon Valley lacks awareness of handling dangerous technology and prioritizing people. His resignation follows Anthropic's Jacob Coxon's September exit, who warned AI could "kill us all" by 2030. Robinson cautioned that OpenAI's "unbounded optimism" means safety failures will grow as systems become more capable.