🧯

Jan Leike

Original nameJan Leike

German Computer Scientist, AI Alignment Researcher

The Math Behind the Algorithm · Theoretical Founders

Co-Leading OpenAI’s "Superalignment" TeamEarly Research on the RLHF MethodAI-Safety Alignment Theory

Who they are

Jan Leike is a German computer scientist, former co-lead of OpenAI’s "superalignment" team (co-led with Ilya Sutskever), a team whose core mission was to research in advance how to ensure that the goals and behavior of the superhuman AI systems that may appear in the future always stay consistent with humanity’s overall interest. In May 2024, not long after Sutskever’s departure, Leike too publicly announced his resignation from OpenAI, writing on social media that the company’s recent internal tendency of "safety culture and process giving way to eye-catching products" was growing ever farther from his own idea of how to responsibly advance the technology, this series of senior-level collective departures provoking wide public attention and discussion at the time over OpenAI’s internal divergence between safety and commercialization. After leaving, Leike joined Anthropic to continue AI-alignment research.

Primary sourcesThe official announcement of OpenAI’s "superalignment" teamLeike’s public resignation statement (2024)

Key stories

"Safety Culture Giving Way to Eye-Catching Products"

On announcing his departure from OpenAI, Leike wrote on social media, frankly saying that within the company in recent years "safety culture and process are giving way to a more eye-catching pace of product releases," and that his team had over the past months found it ever harder to obtain the necessary compute and advance alignment research, this relatively rare public statement letting the world glimpse the tension long present within the world’s most watched AI company between the speed of commercialization and safety caution.

Relationships

Ally

🌌Ilya Sutskever 🛡️Dario Amodei

Echoes today

Below are how modern works borrow or reinterpret this name or story — not the original material. The two differ, so keep them apart.

The public statement that "safety culture gave way to product"Leike’s public remarks on resigning were thereafter widely cited by the media, becoming one of the most-cited first-hand materials for discussing the tension inside an AI company between commercialization and safety caution.

Appears in

The Generative-AI Explosion and the Industry ContestThe Split of Anthropic and "Constitutional AI"

← Back to Figures