AI chip export controls rely on private sector for fast verification
Near-Term Verification Methods for AI Chip Exports
Computers and Society
Summary
Controlling the export of AI chips can influence how advanced AI develops, but checking if these controls are followed is tricky. The authors look at ways to verify chip exports quickly, within about a year, by checking where chips end up, who uses them, and what they are used for. They suggest that the best methods use private companies working with the U.S. government and current technology, instead of relying heavily on government staff. These verification steps could also support future international agreements on controlling AI technology.
AI chipexport controlverification methodsend-location verificationend-user verificationend-use verificationBureau of Industry and Securityprivate sectorregulatory frameworkinternational agreements
Authors
Bruna Avellar, Erich Grunewald
Abstract
AI chip export controls can help the United States shape the development of frontier AI, but their effectiveness depends on reliable methods for verifying compliance. This paper examines near-term verification mechanisms (implementable in approximately one year) and groups them into three categories: end-location verification (whether controlled chips remain in authorized locations and/or jurisdictions), end-user verification (whether entities that acquire or access compute are legitimate), and end-use verification (whether computing power is used for prohibited purposes). We discuss how each mechanism can be implemented within the regulatory framework of the U.S. Bureau of Industry and Security (BIS), outlining implementation steps and identifying which actors can perform verification (BIS, exporters, or accredited third-party auditors). Given BIS's resource constraints, the most viable mechanisms rely on private-sector actors working alongside BIS, leverage existing technologies, and scale without requiring large increases in government staffing. These mechanisms could also help monitor future international agreements on AI.