Paul Christiano

Ricercatore di allineamento che ha guidato il team di allineamento dei modelli linguistici di OpenAI, ha fondato l’Alignment Research Center e dal 2024 ha diretto la sicurezza dell’IA allo US AI Safety Institute; colloca la probabilità di una presa di controllo da parte dell’IA nell’ordine delle decine di punti percentuali.

Ruoli

2017 – 2021Researcher; led the language model alignment team, OpenAICo-authored foundational work on reinforcement learning from human feedback.Fonte →
apr 2021 – oggiFounder and Executive Director, Alignment Research CenterCurrently listed as executive director on his own site, after returning from government.Fonte →
set 2023 – oggiAdvisory board member, UK Frontier AI TaskforceFonte →
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.Fonte →
2026 – oggiTechnical advisor, Center for AI Standards and Innovation, NISTFonte →
9 set 2026 – oggiBoard member, Safety and Security Committee, OpenAI FoundationAppointed to the OpenAI Foundation Board and its Safety and Security Committee, and a non-voting observer on the OpenAI Group PBC Board, while remaining a Senior Technical Advisor at CAISI recused from OpenAI matters; his own statement of the same day (paulfchristiano.substack.com) calls it "the OpenAI nonprofit board".Fonte →

A registro

9 set 2026 · Post · Substack
Based on the recent trajectory of capabilities and the continued difficulty of alignment, I now believe there is a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term. I do not think that the AI industry in general, including OpenAI, is currently on track to reduce this risk to an acceptable level.
Board member, Safety and Security Committee, OpenAI Foundation all’epoca
Written on joining the OpenAI Foundation board and its Safety and Security Committee; footnote 1 puts his all-things-considered risk of catastrophic loss of control at 4% over the next year and 15% over the next three.
Personal statement on joining the OpenAI board →
24 ott 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 all’epoca
Argues that a good responsible scaling policy must name the conditions under which development would be paused.
Thoughts on responsible scaling policies and regulation →
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 all’epoca
Also gives a 20% probability that most humans die within 10 years of building powerful AI.
My views on "doom" →
19 giu 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 all’epoca
First of his listed points of agreement with Yudkowsky.
Where I agree and disagree with Eliezer →