Vigilia.
← The warnings
Extinction riskCivil societyTechnical advisor, Center for AI Standards and Innovation, NIST3 entries on the record

Follow

Paul Christiano

Alignment researcher who led OpenAI’s language-model alignment team, founded the Alignment Research Center, served as head of AI safety at the US AI Safety Institute from 2024, and puts the chance of an AI takeover in the tens of percent.

Roles

  1. 2017 – 2021Researcher; led the language model alignment team, OpenAICo-authored foundational work on reinforcement learning from human feedback.Source →
  2. Apr 2021 – presentFounder and Executive Director, Alignment Research CenterCurrently listed as executive director on his own site, after returning from government.Source →
  3. Sept 2023 – presentAdvisory board member, UK Frontier AI TaskforceSource →
  4. 16 Apr 2024 – 2026Head of AI Safety, US AI Safety Institute, NIST (renamed Center for AI Standards and Innovation, June 2025)Reported in August 2026 to have stepped back to a part-time advisor role in July 2026; the reporting (X) could not be fetched, but his site now lists him as technical advisor.Source →
  5. 2026 – presentTechnical advisor, Center for AI Standards and Innovation, NISTSource →

On the record

  1. 24 Oct 2023 · Post · LessWrong

    I think AI developers are not prepared to work with very powerful AI systems. They don't have the scientific understanding to deploy superhuman AI systems without considerable risk

    Advisory board member, UK Frontier AI Taskforce at the time

    Argues that a good responsible scaling policy must name the conditions under which development would be paused.

    Thoughts on responsible scaling policies and regulation →

  2. 27 Apr 2023 · Post · LessWrong

    Probability of an AI takeover: 22% … Probability that humanity has somehow irreversibly messed up our future within 10 years of building powerful AI: 46%

    Founder and Executive Director, Alignment Research Center at the time

    Also gives a 20% probability that most humans die within 10 years of building powerful AI.

    My views on "doom" →

  3. 19 Jun 2022 · Post · LessWrong

    Powerful AI systems have a good chance of deliberately and irreversibly disempowering humanity. This is a much more likely failure mode than humanity killing ourselves with destructive physical technologies.

    Founder and Executive Director, Alignment Research Center at the time

    First of his listed points of agreement with Yudkowsky.

    Where I agree and disagree with Eliezer →