Will Chinese AI firms' distillation campaigns cause a U.S. lab to publicly disclose proprietary IP theft by year-end 2026?

CISA reports that China-based AI companies are conducting systematic large-scale knowledge distillation campaigns targeting proprietary U.S. AI models. The campaigns extract functionalities and capabilities as core to their development strategy. This item tracks documented public disclosure of successful IP theft resulting from these campaigns.

Resolution Criteria

YES if a U.S.-based AI lab (OpenAI, Anthropic, Google DeepMind, Meta, xAI, etc.) or U.S. government agency (NSF, NIST, DARPA) publicly confirms in an official statement, press release, congressional testimony, or peer-reviewed publication that proprietary model weights, training data, or architectural details were stolen via distillation by Chinese entities before Dec 31, 2026.