Anthropic accused China’s Alibaba of illicitly extracting capabilities from its Claude AI system via “distillation attacks,” alleging nearly 29 million exchanges were conducted through thousands of fraudulent accounts, according to a June 10 letter to Sens. Tim Scott and Elizabeth Warren. The San Francisco firm urged Congress to penalize perpetrators and bolster safeguards to prevent U.S. AI technology from being repackaged by foreign rivals, calling the operation the largest of its kind. Anthropic cited Pentagon assertions linking Alibaba and other major Chinese companies to the military—allegations the firms deny—and noted similar concerns raised by U.S. developers about Chinese distillation efforts. The dispute comes as Anthropic, a leading OpenAI rival, readies a high-profile listing and faces scrutiny over risks tied to its most advanced models.
Related article:
