Thomas Friedman

AI position

Aware of x-risk

Repeatedly writes about catastrophic/loss-of-control style risks (AI-enabled mass hacking of critical infrastructure, models "scheming" in Anthropic-style evaluations, need for U.S.-China guardrails "as important as nuclear arms control") but frames these as geopolitical/security risks to manage via diplomacy rather than as an existential-risk cause he champions per se.

Why influential

Produces reusable arguments and interpretive frames that circulate among policymakers, donors, advocates, academics and other journalists. Primary policy lanes: foreign policy; trade; climate; technology.

Who pays attention

Executives, diplomats, investors, policymakers, academics and international media

Work samples

• The One Danger That Should Unite the U.S. and China

  Sep 2, 2025 · article

  Argues AI's dual-use dangers (empowering 'super-empowered' bad actors) require a joint US-China 'trust architecture,' comparing it to Cold War nuclear arms control; names Anthropic, OpenAI, Google/Alphabet, Meta, Alibaba, DeepSeek, ByteDance as the companies to coordinate.

• Anthropic's Restraint Is a Terrifying Warning Sign

  Apr 7, 2026 · article

  Covers Anthropic's limited-release 'Claude Mythos Preview' and 'Project Glasswing' vulnerability-finding consortium; uses it to again argue for US-China AI cooperation on security guardrails.

• Tom Friedman Thinks We're Getting China Dangerously Wrong

  Apr 15, 2025 · podcast

  Guest appearance on The Ezra Klein Show; discusses AI, EVs, and clean energy advances he observed in China and argues the US-China adversarial framing is outdated.

• Fireside Chat with Thomas Friedman and Craig Mundie

  Jan 20, 2026 · video

  Public Stanford conversation (moderated by John Hennessy) on AI as a new era of civilization and a proposed trust framework for AI systems developed with US/Europe/China input.

AI-world contacts

• Dario Amodei — Co-founder & CEO, Anthropic

  • Wrote aboutdirectness unclassified · Apr 7, 2026 — Column covers Anthropic's restricted release of 'Claude Mythos Preview' and its 'Project Glasswing' initiative, describing the company's internal reasoning and safety trade-offs in depth (organization-level; Amodei not named directly in available excerpts).

• Craig Mundie — Technology strategist / AI policy co-author, Formerly Microsoft; co-author of 'Genesis'

  • Wrote aboutdirectness unclassified — Friedman's recurring AI advisor and public interlocutor; co-wrote 'Genesis: Artificial Intelligence, Hope, and the Human Spirit' with Henry Kissinger and Eric Schmidt, and has done multiple joint public events with Friedman (Stanford, Planet Word Museum) shaping Friedman's AI framing.