America’s adversaries are using top American tech — without telling their customers

Millions of overseas customers are using American-made AI companies, and they don’t even know it.A new report released by Anthropic will likely throw a wrench into what is known about the AI race, pointing to the idea that the United States’ adversaries are not as technologically advanced as they purport to be.’Over 12.1 million exchanges observed.’Anthropic explained in a lengthy report that Chinese companies and other international role players are seemingly reliant, to some degree, on the American company’s AI programs.To that end, multiple Chinese AI companies were discovered to have been forwarding their own customers’ inquiries to Claude and relaying them back to users as if they were their own.Moonshot AI, which uses the chatbot Kimi, was described as “silently” forwarding customer requests to Claude instead of processing them on its own.”Moonshot then displayed Claude’s responses to users. These users thought they were using a Kimi model, but received responses from Claude instead,” Anthropic wrote.Anthropic said that in just a 10-day period, Moonshot relayed almost 300,000 customer requests to the American company, mostly to the AI program Claude Opus. This was done through over 5,000 proxy accounts predominantly located in Singapore and Japan.DeepSeek, another Chinese AI company, was reportedly caught relaying its inquiries to Anthropic’s Claude too, even when it was a company using its product.Anthropic said one Chinese company was “almost certainly not made aware” that its data was being relayed to Claude when it used DeepSeek to analyze internal documents that included sensitive information and organizational structure. RELATED: Hackers are trying to steal money from X users — don’t be one of them‼️ BREAKING: Chinese labs DeepSeek and Moonshot were quietly relaying customer prompts to Claude through fraudulent accounts. Users thought they were using Chinese AI, but were actually getting answers from Claude.Their customers’ sensitive data now is in Anthropic’s hands. One… pic.twitter.com/haVaTSVRol— International Cyber Digest (@IntCyberDigest) September 10, 2026Additionally, requests from an IT operator from a Russian government agency “associated with its Ministry of Defense” were relayed to Anthropic through DeepSeek, which allegedly exposed live credentials for a Russian government database.Furthermore, a municipal Public Security Bureau in China used DeepSeek, which turned out to be Claude, to build a case management system. An engineer reportedly built a tool that compares a person’s movements against police records and used the national ID number of its citizens.”Scale of distillation attacks attributable to DeepSeek over 14 days in July 2026: over 12.1 million exchanges observed,” Anthropic wrote.Another Chinese tech company called Xiaomi was accused of “illicit distillation” — making 400,000 exchanges with Anthropic over the course of 20 days to siphon data to train its own AI models.RELATED: OpenAI’s newest model is all techies want — but how uncontrollable is it? CFOTO/Future Publishing/Getty Images Anthropic also revealed that three Iranian government-aligned accounts were using Claude to “set up influence operations campaigns.”The campaigns had stated goals to shape public opinion at home and abroad and were working on behalf of the Islamic Culture and Communications Organization under the Ministry of Culture and Islamic Guidance and the Islamic Propaganda Office of Khorasan Razavi.The work was a portion of a nine-part “international influence” portfolio and included a complete organizational plan for the funeral of Iran’s supreme leader.Like Blaze News? Bypass the censors, sign up for our newsletters, and get stories like this direct to your inbox. Sign up here!