Jan Leike is a German AI researcher specializing in AI alignment and safety. He earned his PhD at the Australian National University, worked as a researcher at DeepMind, and then led the alignment team at OpenAI, where in 2023 he co-led the Superalignment team with Ilya Sutskever. In 2024 he joined Anthropic to lead its alignment research.
In 2017 he co-authored, with Paul Christiano, Dario Amodei, and others, a paper combining human preference comparisons with deep reinforcement learning, an approach that became the foundation of RLHF.