AI Development Freeze Would Cede Advantage to China, Argues Thiessen

The greatest artificial intelligence threat facing the United States may not be rogue superintelligence or existential catastrophe — but rather falling behind China in AI capabilities due to self-imposed development constraints, according to Washington Post columnist Marc Thiessen.

Writing in an opinion piece published September 17, 2026, Thiessen argues that calls for nuclear-style freezes on AI development would primarily benefit Beijing while hampering American technological leadership.

The extinction debate

The column addresses recent warnings from former Anthropic researcher Jacob Coxon, who declared on X that AI "could kill us all by the end of the decade." Thiessen contrasts this alarm with the assessment of former Meta chief AI scientist Yann LeCun, who characterized the probability of AI-driven extinction as "smaller than the probability of an extinction-level asteroid hitting the earth" and "a hell of a lot less likely than a nuclear Holocaust."

The stark disagreement among AI researchers highlights the uncertainty surrounding long-term AI risks and the challenge of calibrating policy responses to speculative threats.

Why it matters

The debate over AI development constraints carries direct implications for U.S.-China technological competition. Any unilateral American slowdown in AI research would likely leave Chinese development unchecked, potentially shifting the balance of AI capabilities toward an authoritarian regime with fewer safety guardrails and different values around deployment. The framing positions AI policy as fundamentally a question of strategic competition rather than purely technical risk management.

The strategic calculus

Thiessen's argument rests on the premise that China would not honor or participate in voluntary AI development pauses, making such restraints asymmetric in their impact. This creates a prisoner's dilemma where the nation that proceeds cautiously risks technological subordination to a competitor proceeding at full speed.

The column implicitly challenges the logic of AI safety advocates who have called for development moratoriums, suggesting that geopolitical realities make such proposals strategically untenable regardless of their technical merits.

The piece reflects broader tensions in AI policy circles between those prioritizing existential risk mitigation and those emphasizing the national security imperative of maintaining AI leadership. These competing frameworks increasingly shape regulatory debates and research funding priorities.

The details were first reported by Marc Thiessen in The Washington Post.