US HB8283: Deterring American AI Model Theft Act of 2026

In one paragraph

The federal H.R. 8283 (Deterring American AI Model Theft Act of 2026) would direct the Secretary of State to assess model extraction attacks on closed-source American AI models by entities of concern and to keep a public attackers list, and would direct Commerce to consider Entity List additions. It would authorize IEEPA sanctions. Referred to the House Foreign Affairs Committee April 15, 2026; ordered reported as amended 43-0 April 22, 2026; no reported text yet; it has not passed the House.

Plain-English summary, written with AI assistance. The official text controls.

What it means for business

No duties for U.S. AI model owners; sharing information with Commerce would be voluntary. Foreign entities found to extract models or sell access accounts could face Entity List review and sanctions.

Key provisions

1. 1Defines model extraction attacks as unauthorized extraction of a closed-source AI model's capabilities to replicate or improve another model, such as by circumventing access controls or using fraudulent credentials (Sec. 3)
2. 2Secretary of State, with export-policy agencies, must assess within 180 days of enactment which entities of concern conduct such attacks or are fraudulent account network providers (Sec. 4)
3. 3Commerce must report findings within 210 days, with annual updates for 3 years, publish best practices, and set up voluntary confidential information sharing with model owners
4. 4State must keep a public 'AI Model Extraction Attackers List' for up to 5 years
5. 5Within 210 days Commerce's Under Secretary for Industry and Security must decide whether identified entities should be added to the Entity List (Sec. 5(a))
6. 6The President may block property of identified entities of concern under IEEPA, with the IEEPA penalties applying (Sec. 5(b))

Latest legislative action

Apr 22, 2026. Ordered to be Reported (Amended) by the Yeas and Nays: 43 - 0.

Sponsors (showing 5 of 13)