IC Technology — 2026-04-23

White House OSTP accuses China of industrial-scale AI distillation campaigns targeting U.S. frontier models

The White House Office of Science and Technology Policy released a memorandum on April 23 accusing China of conducting industrial-scale campaigns to distill U.S. frontier AI systems, using tens of thousands of proxy accounts to evade detection and jailbreaking techniques to expose proprietary information. OSTP Director Michael Kratsios wrote that these distillation campaigns also strip security protocols and neutrality mechanisms from the resulting models. The administration announced plans to share threat information with U.S. AI companies and explore measures to hold foreign actors accountable, ahead of a planned Trump-Xi summit.

Analysis
The OSTP memorandum, corroborated independently by CNBC and Fox Business, with CNN and U.S. News as secondary rewrites of FT's original, converts proxy-account-mediated model distillation into a policy instrument by framing commercial API access as a national security vulnerability. New enforcement targeting Chinese distillation campaigns is unlikely within 90 days, a judgment downgraded from likely by weighting the structural mismatch between export-control authority (which governs tangible goods and software, not access patterns) and the absence of named precursors such as a draft BIS rule or Federal Register ANPRM. The announced threat-sharing with U.S. AI companies signals industry coordination as the near-term path, a lower-friction response that sidesteps the rulemaking timeline. The memorandum may be primarily a negotiating chip for the planned Trump-Xi summit, elevating a commercially inconvenient practice to state-sponsored theft rather than establishing a genuine enforcement runway.

4 sources