Bessent Blames Chinese AI for French Grocery Attack

The Treasury Secretary's claims about a refrigeration sabotage operation have not been independently confirmed.

Updated on Nov. 20, 2026 in Artificial Intelligence

Bold flat-color editorial illustration of a stone institutional facade with a lone pale blue refrigeration unit at its base, evoking an unconfirmed government accusation.
Treasury Secretary Scott Bessent on November 12 blamed a Chinese AI model for disabling refrigeration in up to 700 French grocery stores, a claim no independent source has confirmed. AI Illustration. Upload story photo >

Live Poll

Should the US take a harder line against Chinese AI companies?

US Treasury Secretary Scott Bessent said on November 12-13, 2026 that a Chinese AI model was used to disable refrigeration systems in 150 to 700 French grocery stores, attributing the operation to Iranian operatives. No independent verification has confirmed the connections Bessent drew between the attack and a Chinese model.

Why it matters

The claim links Chinese AI models to a physical attack, raising the stakes of US scrutiny of how those models were built. It comes as US agencies accuse Chinese firms of harvesting American AI outputs at scale.

Distillation trains a smaller model on the outputs of a more capable one, and US agencies accused six Chinese firms of doing this at scale against American models going back to late 2024. The technical extent of any model carrying traces of US outputs has not been publicly demonstrated.

The players

Scott Bessent

Scott Bessent is the US Treasury Secretary, a cabinet-level official who oversees the Treasury Department and has spoken publicly on economic and national security matters.

Moonshot AI

Moonshot AI is a Chinese artificial intelligence company that develops the Kimi family of large language models.

Anthropic

Anthropic is a US artificial intelligence company that develops the Claude family of AI models, including the Fable model referenced in the accusations.

The details

Bessent said US authorities would scrutinize Chinese AI models for potential intellectual property theft, referring to a practice he called industrial-scale distillation. He suggested Chinese models such as Moonshot AI's Kimi K3 could carry watermarks revealing training on outputs from US models, and singled out Kimi K3 in connection with large-scale distillation from Anthropic's Fable model.

Timeline

  1. July 21-22, 2026: Bessent said the US would scrutinize Chinese AI models for IP theft.

  2. September 2026: US agencies accused six Chinese firms of distillation attacks on American models.

  3. September 2026: US officials proposed a hotline with China to manage AI risks.

  4. November 12-13, 2026: Bessent made the claim about the French grocery store refrigeration attack.

The Tech Race

The accusations extend a US-China rivalry over frontier AI from research and chips into the question of who can absorb American model capabilities without paying for them. The September 2026 accusations against six Chinese firms, alongside a proposed US-China AI risk hotline, show Washington pursuing enforcement and dialogue at once.

If US action moves from scrutiny to sanctions or Entity List designations, Chinese AI tools could become harder for American businesses and users to access through partners and suppliers. Firms that rely on these models may face new compliance questions about where their tools were trained.

The takeaway

Claims that tie AI models to physical attacks remain unverified, so readers should weigh official accusations against independent evidence as they emerge. Watch for whether US agencies publish technical findings on distillation.

Further reading

For broader context, see our coverage of Artificial Intelligence developments and policy.

Source note: This article includes information reported by Crypto Briefing.

Live Poll

Should the US take a harder line against Chinese AI companies?