PeopleofSuperintelligence

Lab leader

Paul Christiano

Paul Christiano is an AI researcher and lab director whose work connected human feedback with language-model training. He earned a mathematics degree at MIT and a computer-science PhD at UC Berkeley before leading OpenAI’s language-model alignment team and co-authoring InstructGPT. He founded the Alignment Research Center (ARC) for theoretical alignment research, and TIME included him in its 2023 list of the most influential people in AI. In April 2024, he became head of AI safety at the United States AI Safety Institute. As of October 2026, he is ARC’s executive director and a technical advisor at the Center for AI Standards and Innovation within NIST; he also serves on OpenAI’s Foundation Board and Safety and Security Committee.

Now at
OpenAI, Alignment Research Center and US AI Safety Institute / CAISI
Works
30

Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Photo of Paul Christiano

Timeline

  1. 2012–2017

    Mathematics and learning theory

    He graduated from MIT with a mathematics degree in 2012. At UC Berkeley, he completed a computer-science PhD in 2017 with a dissertation on manipulation-resistant online learning.

  2. 2017–2021

    Training on human preferences

    From 2017 to 2021, he led OpenAI’s language-model alignment team and worked on reinforcement learning from human feedback. His 2017 paper, Deep reinforcement learning from human preferences, trained agents using human preferences between pairs of trajectory segments.

  3. 2021

    Opening an alignment center

    He left OpenAI in 2021 to focus on theoretical AI alignment and founded the Alignment Research Center (ARC).

  4. 2022

    Teaching models to follow instructions

    He co-authored Training language models to follow instructions with human feedback. The paper introduced InstructGPT and fine-tuned GPT-3 with labeler demonstrations and ranked model outputs, followed by reinforcement learning from human feedback.

  5. 2023

    A place on TIME100 AI

    TIME included him in its 2023 list of the most influential people in AI.

  6. April 2024

    Taking AI safety into government

    He became head of AI safety at the United States AI Safety Institute at NIST in April 2024.

  7. September 2026

    Joining OpenAI’s foundation board

    He joined OpenAI’s Foundation Board and its Safety and Security Committee.

  8. As of October 2026

    Leading ARC and advising NIST

    He is ARC’s executive director and a technical advisor at the Center for AI Standards and Innovation within NIST.

Sources and credits

Photo: nist.gov; Paul Christiano headshot; creator not stated. License: Not stated by source. Source: nist.gov. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. MIT ESP - Paul Christiano esp.mit.edu. Checked 14 October 2026.
  2. Manipulation-resistant online learning | EECS at UC Berkeley www2.eecs.berkeley.edu. Checked 14 October 2026.
  3. Paul Christiano simons.berkeley.edu. Checked 14 October 2026.
  4. Paul Christiano - Wikipedia en.wikipedia.org. Checked 14 October 2026.
  5. Deep reinforcement learning from human preferences arxiv.org. Checked 14 October 2026.
  6. Training language models to follow instructions with human feedback arxiv.org. Checked 14 October 2026.
  7. U.S. Commerce Secretary Gina Raimondo Announces Expansion of U.S. AI Safety Institute Leadership Team nist.gov. Checked 14 October 2026.
  8. Paul Christiano paulfchristiano.com. Checked 14 October 2026.
  9. Team — Alignment Research Center alignment.org. Checked 14 October 2026.
  10. OpenAI adds a prominent AI doomer to its board of directors | TechCrunch techcrunch.com. Checked 14 October 2026.