OPENAI’S TOP SAFETY EXPERT QUITS. THE EXITS KEEP COMING. Aleksander Madry, the computer scientist who joined OpenAI in 2024 as head of preparedness, has announced he is leaving the company. Madry spent nearly three years at OpenAI, first leading the preparedness team — the group responsible for evaluating catastrophic risks from frontier models — before being reassigned to an AI reasoning role last summer. Now he says he wants to focus on AI’s impact on the economy and start something new in that space. The timing is not lost on anyone paying attention. Madry’s departure comes days after Andrej Karpathy, an OpenAI co-founder and one of the most respected researchers in the field, announced he was leaving to join Anthropic. The stream of senior safety and technical talent flowing out of OpenAI has become a pattern, not an anomaly. In 2024, the entire Superalignment team dissolved. Paul Christiano. Jan Leike. Ilya Sutskever. Each departure was explained away individually. Together they tell a different story. OpenAI is preparing an IPO targeting a September market debut at an $850 billion valuation. That commercial pressure and the demands of scaling a frontier operation appear fundamentally at odds with the patient, research-intensive culture that safety work requires. Madry did not publicly blame the company in his announcement. He did not need to. The pattern has already spoken. Keywords: OpenAI safety exodus, Aleksander Madry departure, OpenAI preparedness team, AI safety researcher leaves OpenAI