Jack Clark

Confidence 0.75 · 2 sources · last confirmed 2026-05-08

Co-founder of Anthropic and Head of Public Benefit (2026, leading the newly-launched Anthropic Institute). Earlier role at Anthropic: Head of Policy. One of the most public-facing AI-policy practitioners across the wiki’s responsible-ai and enterprise-ai-adoption threads — most of his surface in the public record is governance / evaluation / measurement-coloured rather than systems-coloured.

Career trajectory: technical journalist (Bloomberg BusinessWeek, The Register) → Strategy and Communications Director, then Policy Director, at OpenAI (until late 2020) → co-founded Anthropic with Dario and Daniela Amodei and several other ex-OpenAI colleagues (2021). Author of the long-running Import AI weekly newsletter, which is a primary-source channel for many of his policy positions.

Public-policy roles relevant to wiki sources:

  • Founding member of the AI Index at Stanford HAI (2017–2024); continues on the Steering Committee for the 2026 edition (representing Anthropic / OECD per the AI Index page’s member list).
  • Inaugural member, U.S. National AI Advisory Committee (NAIAC) — 2021–2024 term, alongside James Manyika (NAIAC vice-chair).
  • OECD AI Group of Experts — co-chair / member.
  • Advisory roles at CNAS and other AI-policy organizations.

Recent published positions (2026): blog posts forecasting that AI R&D will be fully automated by 2028, with 60%+ odds that AI systems will train their own successors before 2029. These positions place him firmly in the short-timeline / high-capability-trajectory camp of the AI-safety community — useful context for reading Anthropic’s policy and engineering output (e.g. 2026-05-07-anthropic-managed-agents-decoupling-brain-hands and 2026-05-07-anthropic-economic-index-5-learning-curves) where the timeline assumption is implicit.

Appears in this wiki via