American artificial intelligence company Anthropic has accused Chinese technology giant Alibaba’s Qwen AI lab of carrying out what it described as the largest known “adversarial distillation attack” against its Claude AI models. The allegations, made in a formal letter to US lawmakers, have intensified concerns over AI intellectual property theft and the growing technological rivalry between the United States and China.
Ahead of a US Senate hearing on artificial intelligence, Anthropic sent a letter dated June 10 to US Senators Tim Scott and Elizabeth Warren of the Senate Committee on Banking, Housing, and Urban Affairs. In the letter, the company alleged that Alibaba’s AI division orchestrated millions of AI distillation attacks aimed at extracting the capabilities of its Claude models.
According to Anthropic, operators linked to Alibaba and its Qwen AI lab allegedly created around 25,000 fraudulent accounts that generated more than 28.8 million interactions with Claude over a six-week period between April 22 and June 5, 2026. The company claimed these exchanges were conducted to “brazenly and illicitly” copy Claude’s capabilities, calling it the largest distillation campaign it has detected to date.
What is AI distillation?
AI model distillation is a machine learning technique in which a smaller or less capable model is trained using responses generated by a more advanced AI system. In an adversarial distillation attack, developers repeatedly query a powerful model to collect high-quality outputs, which are then used to improve another AI model without obtaining direct access to the original model’s architecture or internal parameters.
According to Anthropic, this approach enables competing AI developers to replicate sophisticated features such as reasoning, coding and agentic behaviour while significantly reducing the cost of developing advanced large language models.
The company said AI distillation is increasingly being adopted as a cost-effective strategy by competitors, particularly those affected by restrictions on advanced AI chips and computing resources, allowing them to narrow the gap with leading American AI companies, including Anthropic, OpenAI and Google.
Claude’s advanced capabilities allegedly targeted
Anthropic alleged that the campaign specifically sought to extract Claude’s most commercially valuable features, including software engineering, coding, agentic reasoning, autonomous decision-making and task execution capabilities. The company also claimed its long-horizon reasoning and ability to perform complex tasks were among the primary targets.
Also Read:Andy Jassy Meets PM Modi: Amazon Pledges $48 Billion Investment In India By 2030
The company further alleged that the activities violated its terms of service as well as geographic access restrictions, noting that Claude is not officially available in China. It also claimed the attacks ignored repeated warnings from the US government regarding AI intellectual property theft.
Previous AI distillation campaigns
Anthropic said this is not the first time it has detected similar activities. In February 2026, the company publicly disclosed smaller-scale AI distillation campaigns allegedly linked to other Chinese AI laboratories.
According to Anthropic, it recorded approximately 150,000 exchanges from DeepSeek, more than 3.4 million exchanges from Moonshot AI and around 13 million exchanges from MiniMax. However, the company said these campaigns were significantly smaller than Alibaba’s alleged 28.8 million interactions.
US scrutiny intensifies
Shortly after Anthropic submitted its letter, the US Department of Commerce imposed restrictions on access to the company’s latest AI models, Mythos and Fable, on June 12. The department cited concerns that the models could potentially be used by military and intelligence entities in China and other countries of concern. Anthropic later disabled global access to both models.
Around the same period, Alibaba was added to the US Department of Defense’s list of Chinese military companies alongside several other Chinese firms. Alibaba is contesting the designation.
Anthropic has urged closer cooperation between the government and the AI industry, stronger intelligence sharing and tougher penalties to safeguard American leadership in artificial intelligence. The company also reaffirmed its support for US government efforts to prevent AI intellectual property theft and said it remains committed to working with Congress and the administration on the issue.
The allegations have also had a market impact, with Alibaba’s shares reportedly falling about 3 per cent following reports of the complaint. The controversy has further fuelled debate over AI security, intellectual property protection and the escalating global race to dominate artificial intelligence.
-
-
-
-
