The U.S. Department of State has mandated a worldwide initiative aimed at drawing attention to what it alleges are extensive efforts by Chinese firms, among them the artificial intelligence startup DeepSeek, to pilfer intellectual property from American AI research facilities.
This information was revealed in a diplomatic cable reviewed by the Reuters news agency.
The communique, circulated to diplomatic and consular posts globally, instructs personnel at these missions to engage their foreign counterparts in discussions concerning “concerns regarding China’s distillation of U.S. AI models.”
“A separate inquiry and message has been transmitted to Beijing for discussion with China,” the document states.
Distillation refers to the methodology employed to train smaller AI models by leveraging the output generated by larger, more resource-intensive models, as a means to cut the expenses associated with developing a potent new AI instrument.
It was reported back in February that OpenAI had alerted U.S. lawmakers that DeepSeek was targeting the developer of ChatGPT and the nation’s top AI enterprises with the objective of duplicating models for utilization in their own training regimes.
Beijing rejects these accusations.
On Friday, the Chinese Embassy in Washington reiterated its stance that the allegations are without merit.
“Claims that Chinese companies are infringing upon American AI intellectual property are unfounded and constitute deliberate attacks against China’s advancement and progress in the AI sector,” a statement provided to Reuters read.
This declaration followed shortly after the Chinese startup DeepSeek unveiled a preview of its recently updated model, a development we previously covered.
About the Author
