The U.S. State Department has circulated a directive to diplomatic and consular missions across the globe urging them to raise alarms over what Washington describes as widespread efforts by Chinese companies to appropriate intellectual property from U.S. artificial intelligence research, according to a diplomatic cable reviewed by Reuters.
The cable, dated Friday, specifically cites the Chinese AI startup DeepSeek among other firms and sets out an objective 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."
Distillation, as the cable explains, refers to the practice of training smaller AI models by using the outputs of larger, more resource-intensive models. The technique can reduce the cost of creating new AI tools by leveraging the behavior of established systems, but the State Department document contends such campaigns can result in models that appear to match benchmark performance at lower cost while failing to replicate the full capabilities of the originals.
In its guidance, the State Department warned that distillation efforts by foreign actors can remove security safeguards and undo mechanisms designed to make models ideologically neutral and truth-seeking. The cable instructs diplomats to engage with their foreign counterparts about "concerns over adversaries' extraction and distillation of U.S. A.I. models." It also notes that a separate demarche and message have been sent to Beijing to raise these issues directly with China.
The memo names specific Chinese firms including DeepSeek, Moonshot AI and MiniMax. DeepSeek, which drew widespread attention last year for a low-cost model, this week unveiled a preview of a new model adapted for Huawei chip technology, a development the cable highlights as evidence of growing technology independence in China.
The State Department, DeepSeek and the Chinese Embassy in Washington did not immediately respond to requests for comment. Representatives for Moonshot AI and MiniMax also did not immediately reply.
Earlier this week, the White House made related accusations about Chinese efforts to replicate U.S. AI systems. The Chinese Embassy in Washington dismissed those claims as "baseless allegations," saying Beijing "attaches great importance to the protection of intellectual property rights."
The State Department cable, which has not been publicly reported before, signals that the current administration is treating worries about foreign distillation of U.S. AI models as a substantive policy concern. The cable states that models born from "surreptitious, unauthorized distillation campaigns" can produce products that perform comparably on select benchmarks at a fraction of the cost but do not replicate the original system's full performance. It adds that such campaigns also "deliberately strip security protocols from the resulting models and undo mechanisms that ensure those AI models are ideologically neutral and truth-seeking."
OpenAI has previously warned U.S. lawmakers that the Chinese startup DeepSeek was targeting the ChatGPT maker and other leading U.S. AI companies with the aim of reproducing models and using them for its own training, according to reporting in February cited by the cable.
The cable and a follow-up memo arrive just weeks before U.S. President Donald Trump is scheduled to travel to Beijing for meetings with Chinese President Xi Jinping, a timing that could add friction to an already tense technology rivalry between the two countries despite a recent detente.
Which stock should you buy in your very next trade? AI computing powers are changing the stock market. Investing.com's ProPicks AI includes dozens of winning stock portfolios chosen by an advanced AI. Our flagship Tech Titans strategy doubled the S&P 500 within 18 months, including notable winners like Super Micro Computer (+185%) and AppLovin (+157%). Which stock will be the next to soar? Pick Stocks with AI.