AI Espionage: U.S. Accuses China of Distillation Tactics

The U.S. government has accused Chinese AI companies of unethical practices, claiming they copied American technology through 'distillation'. This accusation comes before meetings between the two nations' leaders. The U.S. alleges these actions might be state-approved and could bolster China's military capabilities.

The United States has leveled serious allegations against Chinese AI enterprises, asserting that these companies have maliciously duplicated technology from American firms. This development surfaces just before a crucial meeting between the leaders of both countries, potentially impacting diplomatic relations.

According to U.S. officials, Chinese companies like DeepSeek, Moonshot AI, and Alibaba have employed a technique known as distillation to train smaller AI models using data from larger American AI models, including those from Anthropic, OpenAI, and Google's parent company Alphabet. The officials suggest that these actions were likely executed with knowledge from the Chinese government.

The timing of these revelations aligns with an impending visit from Chinese President Xi Jinping and a U.S.-China dialogue focused on AI safety risks. The issue has stirred concerns about national security, as it is believed that the replicated technology could bolster China's military and cyber capabilities, posing a direct threat to the U.S. and its allies.