US and China Trade Accusations Over AI Theft Claims
▼ Summary
– The US is preparing to crack down on China’s alleged “industrial-scale theft” of American AI labs’ intellectual property.
– Since DeepSeek’s launch, OpenAI and other firms have accused global rivals of using “distillation” to steal AI model outputs.
– Google claimed actors attempted to clone its Gemini chatbot by prompting it over 100,000 times to train cheaper copycats.
– Anthropic accused Chinese firms DeepSeek, Moonshot, and MiniMax of generating over 16 million exchanges with Claude via 24,000 fraudulent accounts.
– White House official Michael Kratsios warned that foreign entities, principally in China, are engaged in deliberate campaigns to distill US frontier AI systems.
The United States is gearing up to confront what it describes as China’s systematic and large-scale theft of intellectual property from American AI labs, according to a Thursday report from the Financial Times. This escalation follows the emergence of DeepSeek, a Chinese model that OpenAI has alleged was built using outputs from its own systems. Since then, a growing number of AI companies have accused global competitors of leveraging a technique known as distillation to unlawfully replicate their technology.
In January, Google reported that commercially motivated actors, not exclusively based in China, had attempted to clone its Gemini AI chatbot. The company said these actors prompted the model over 100,000 times in an effort to train cheaper imitations. The following month, Anthropic leveled accusations against Chinese firms DeepSeek, Moonshot, and MiniMax, claiming they used the same distillation method to generate more than 16 million exchanges with its Claude model through roughly 24,000 fraudulent accounts. Also in February, OpenAI confirmed that the majority of such attacks it observed originated from China.
From Washington’s perspective, these distillation campaigns pose a significant threat, potentially allowing China to rapidly close the gap in the global AI race. A memo reviewed by the FT, authored by White House Office of Science and Technology Policy Director Michael Kratsios, stated bluntly that the U. S. government has evidence of foreign entities, primarily based in China, engaging in deliberate, industrial-scale efforts to distill American frontier AI systems.
(Source: Ars Technica)
