According to a diplomatic cable reviewed by Reuters, the U.S. State Department has authorized a global push to draw attention to what it claims are pervasive attempts by Chinese companies, including AI company DeepSeek, to steal intellectual property from U.S. AI labs.
The letter advises diplomatic personnel to discuss “concerns over adversaries’ extraction and distillation of U.S. A.I. models” with their foreign counterparts. It was delivered on Friday to diplomatic and consular facilities worldwide.
“A separate demarche request and message has been sent to Beijing for raising with China,” the document states.
In an attempt to reduce the cost of training a potent new AI tool, distillation is the process of training smaller AI models using the output from larger, more expensive ones.
Similar charges were made by the White House this week, although no prior reports of the cable have been made. A request for comment was not immediately answered by the State Department.
According to a February Reuters story, OpenAI alerted U.S. lawmakers that DeepSeek was pursuing the nation’s top AI firms and the manufacturer of ChatGPT in order to duplicate models and utilize them for its own training.
CHINA REJECTS ACCUSATIONS
The Chinese Embassy in Washington on Friday reiterated its stance that the accusations are baseless.
“The allegations that Chinese entities are stealing American AI intellectual property are groundless and are deliberate attacks on China’s development and progress in the AI industry,” it said in a statement to Reuters.
China’s increasing independence in the field was shown on Friday when DeepSeek, whose low-cost AI model startled the globe last year, unveiled a sneak peek of a highly anticipated new model, named the V4, customized for Huawei chip technology.
A request for comment was also not quickly answered by DeepSeek. It has previously said that its V3 model did not purposefully employ artificial data produced by OpenAI; instead, it used data that was organically occurring and gathered through web crawling.
Citing worries over data privacy, several Western and several Asian governments have prohibited their institutions and personnel from utilizing DeepSeek. However, on global platforms that host open-source models, DeepSeek’s models have continuously been among the most popular.
The State Department cable said its purpose was to “warn of the risks of utilizing AI models distilled from U.S. proprietary AI models, and lay the groundwork for potential follow-up and outreach by the U.S. government.”
It also mentioned Chinese AI firms Moonshot AI and MiniMax. Neither company immediately responded to a request for comment.
The cable said that “AI models developed from surreptitious, unauthorized distillation campaigns enable foreign actors to release products that appear to perform comparably on select benchmarks at a fraction of the cost but do not replicate the full performance of the original system.”
Furthermore, the efforts “deliberately strip security protocols from the resulting models and undo mechanisms that ensure those AI models are ideologically neutral and truth-seeking.”
Just a few weeks before U.S. President Donald Trump is scheduled to meet with Chinese President Xi Jinping in Beijing, the White House allegations and the cable were made. They might intensify the two countries’ long-running tech battle, which had been defused by a detente mediated in October of last year.
