Eliezer Yudkowsky is a self-taught AI researcher and co-founder of the Machine Intelligence Research Institute (MIRI). He is among the earliest and most vocal advocates for treating advanced AI as an existential risk, and his "Sequences" essays, published on LessWrong, are credited with shaping much of the modern AI safety and rationalist community.
Background
Yudkowsky is a self-taught AI researcher with no formal academic affiliation. His "Sequences" essays on LessWrong laid out the intellectual foundations later drawn on by the AI safety and rationalist communities.
Roles
Yudkowsky co-founded MIRI and has served as a long-time research fellow there.
Positions and statements
Yudkowsky's arguments center on the claim that superintelligent AI systems are likely to be misaligned with human values by default, and that current alignment research is insufficient to prevent catastrophe. He treats Deceptive Alignment as a core element of his threat model, a framing that has carried into safety discourse across labs and policy circles. Critics contend his positions overweight speculative risk relative to present harms.
In 2023, Yudkowsky co-signed and amplified the FLI open letter calling for a pause on giant AI experiments, and argued beyond the letter's terms for an outright moratorium on frontier AI development.
Relationships
- leads: Machine Intelligence Research Institute (MIRI) — co-founder and long-time research fellow
- supports: FLI — Pause Giant AI Experiments: An Open Letter — signatory and advocate for development moratorium
- depends-on: Deceptive Alignment — central to his threat model