US Treasury Secretary Scott Bessent said on Tuesday that the US will investigate whether Chinese artificial intelligence models were developed by ‘distilling’ American models, and he did not rule out the possibility of imposing sanctions for such practices. He emphasized that the Trump administration does not support ‘theft of intellectual property.’
Chinese open-weight AI models are gaining traction compared to leading offerings from American companies such as OpenAI and Anthropic. This trend raises concerns among tech executives and government officials about the sustainability of US leadership in artificial intelligence. Earlier this month, Chinese startup Moonshot AI released a new model called Kimi K3, which outperforms models from American developers, including OpenAI and Anthropic, on certain industry metrics.
‘If we see foreign models stealing from our major companies, we have the ability to impose sanctions on them for this theft,’ Bessent told Fox Business on Tuesday. He explained that the technical term for such a practice is ‘distillation,’ a method of AI training where a smaller model is trained on the outputs of a more powerful model.
Last month, Anthropic sent a letter to the US Senate Committee on Banking, Housing, and Urban Affairs, alleging that Chinese technology company Alibaba had conducted the ‘largest known distillation attack’ against it to date.
Future negotiations and ‘watermarks’
According to a Reuters report on Tuesday, the US and China plan to hold talks on AI in September, with Bessent representing the US. ‘We are finding watermarks of our major US language models on many Chinese models, and this is unacceptable,’ Bessent stated, adding that the US ‘will consider this issue in the coming days or weeks.’
Accusations against American companies
However, Anthropic and OpenAI themselves have faced accusations of illegal use of intellectual property in recent years. In September, Anthropic agreed to pay $1.5 billion to settle a class-action lawsuit with a group of authors who claimed the company illegally downloaded books from pirate databases. In 2023, The New York Times sued OpenAI and Microsoft, alleging that the companies violated copyright by using its intellectual property to train their models.
Source: CNBC
