A fresh dispute has emerged in the artificial intelligence industry after U.S.-based AI company Anthropic accused Chinese technology giant Alibaba of carrying out a large-scale operation to extract the capabilities of its Claude AI models. According to a letter reviewed by Reuters, Anthropic claims the campaign was the largest known attempt to copy its AI technology and involved millions of interactions with its chatbot over several weeks.
The allegations add to growing concerns surrounding AI model security as companies race to build more powerful language models. They also arrive at a time when Washington has been tightening restrictions on advanced AI technologies amid rising geopolitical tensions with China.
Anthropic claims millions of Claude AI conversations were generated
According to the letter, Anthropic alleges that operators affiliated with Alibaba and its AI research division, Qwen, conducted a large-scale “distillation” campaign between April 22 and June 5, 2026. The company claims the operation generated more than 28.8 million conversations with Claude using nearly 25,000 fraudulent user accounts.
Anthropic describes distillation as a process where developers train a smaller or less capable AI model by learning from the responses of a more advanced one. While the technique is commonly used within organizations on models they own or have permission to use, companies have increasingly raised concerns about unauthorized distillation using commercial AI systems, arguing that it could allow competitors to replicate years of research without developing the technology independently.
In its letter, Anthropic claimed the alleged campaign was intended to accelerate China’s ability to match the capabilities of its advanced Mythos Preview models. Reuters reported that the letter was sent on June 10 to U.S. Senators Tim Scott and Elizabeth Warren ahead of a Senate Banking Committee hearing focused on artificial intelligence.
Alibaba had not publicly responded to the allegations at the time Reuters published its report.
AI security concerns continue to grow
This is not the first time Anthropic has raised concerns about suspected AI model extraction. Earlier this year, the company disclosed that it had identified similar campaigns allegedly involving Chinese AI firms, including DeepSeek, Moonshot AI and MiniMax.
According to Anthropic, those earlier operations were significantly smaller than the latest allegations against Alibaba. The company said DeepSeek generated more than 150,000 exchanges with Claude, while Moonshot AI and MiniMax were allegedly responsible for approximately 3.4 million and over 13 million interactions respectively.
Anthropic warned at the time that these activities were becoming increasingly sophisticated and argued that preventing unauthorized AI model extraction would require closer cooperation between technology companies, governments and cybersecurity experts. The company has also expressed support for stronger intelligence-sharing initiatives between private AI developers and U.S. government agencies to help identify and respond to future attacks more quickly.
The allegations come amid heightened scrutiny of China’s artificial intelligence sector. In April, the White House accused China of systematically targeting the intellectual property of American AI companies, claiming such activities were taking place on an industrial scale. While those accusations were not directed specifically at Alibaba, they reflected broader concerns about protecting advanced AI technologies from unauthorized access.
AI rivalry between the US and China continues to intensify
The reported claims arrive during a period of increasing regulatory action surrounding artificial intelligence. Earlier this month, Alibaba was added to the Pentagon’s list of Chinese military companies, a designation the company is currently challenging. Inclusion on the list does not automatically impose sanctions but can affect business relationships and government procurement decisions.
At the same time, the U.S. Commerce Department has reportedly delayed placing DeepSeek on a trade blacklist despite an interagency committee identifying potential national security concerns. According to Reuters, officials are attempting to avoid further escalating tensions between Washington and Beijing while continuing to evaluate the broader impact of AI export controls.
Only days after Anthropic submitted its letter to lawmakers, the Commerce Department introduced new restrictions on the company’s latest Mythos and Fable AI models. The agency reportedly cited concerns that the advanced systems could be accessed by military or intelligence organizations in China and other countries considered sensitive from a national security perspective. Following those restrictions, Anthropic disabled global access to the affected models.
The dispute highlights how artificial intelligence competition is increasingly extending beyond product launches and technological breakthroughs into questions of intellectual property, cybersecurity and national security. As companies invest billions of dollars to develop increasingly capable AI systems, protecting proprietary models is becoming just as important as building them. Whether Anthropic’s latest allegations lead to regulatory action or further responses from Alibaba remains to be seen, but the incident underscores the growing pressure facing AI companies as the global race for advanced artificial intelligence accelerates.
