Aleksander Madry has returned to OpenAI to focus on AI safety and security. The company announced his return on October 19, 2026, five months after his departure in May 2026.
Madry will collaborate with chief scientist Jakub Pachocki. Their work will address the rapid advancement of AI and its implications for both OpenAI's internal strategy and the broader industry.
Before leaving in May 2026, Madry spent nearly three years at OpenAI. He previously led the Preparedness team, which was established in 2023 to identify catastrophic risks in frontier AI, including cybersecurity and biological threats.
In mid-2024, Madry transitioned from the Preparedness team to a role focused on AI reasoning. He remains on leave from his faculty position at MIT as of 2026.
