4 min readJun 27, 2026 09:28 AM IST
First published on: Jun 26, 2026 at 12:27 PM IST
Anthropic has accused Chinese technology giant Alibaba of carrying out the largest known attempt to clone its Claude AI model, alleging the company used nearly 25,000 fraudulent accounts to extract millions of interactions from the chatbot. The allegations come as competition between the US and China over advanced artificial intelligence intensifies.
According to a confidential letter sent by Anthropic to US Senators Tim Scott and Elizabeth Warren ahead of a Senate hearing on AI, operators linked to Alibaba and its AI division, Qwen, allegedly generated more than 28.8 million exchanges with Claude between April 22 and June 5.
Anthropic claims the activity violated its terms of service and was aimed at extracting some of Claude’s most advanced capabilities, including agentic reasoning, software engineering, and long-horizon task execution.
‘Campaign relied on proxy networks’
The company also alleged that the campaign relied on proxy networks and obfuscation techniques to avoid detection.
The AI firm described the incident as the largest distillation attack it has ever detected. Distillation refers to the process of querying an advanced AI model to reproduce its capabilities in another system without incurring the cost of developing a frontier model from scratch.
Anthropic said the campaign took place even after US President Donald Trump issued a directive in April warning against what he described as “industrial-scale” AI theft by Chinese firms. The company alleged that Alibaba continued its efforts despite the warning and argued that such attacks effectively turn billions of dollars of American AI research into an advantage for geopolitical rivals.
Alibaba already on US blacklist
Alibaba has not publicly responded to the allegations. The company is already involved in a separate legal dispute with the Trump administration after challenging its inclusion on a US blacklist over alleged links to the Chinese military. Alibaba has denied the accusations, stating that it operates independently and focuses on commercial technologies rather than defence or intelligence.
In its letter, Anthropic urged Congress to introduce stronger measures to combat AI model cloning. The company proposed updating antitrust laws to allow AI developers to share information about emerging threats, tightening export controls on advanced AI chips, and introducing penalties that would restrict Chinese companies’ access to US AI models, semiconductors, and overseas data centres if they are found to be engaging in illicit distillation.
Anthropic argued that slowing China’s progress in frontier AI would give the US more time to strengthen cybersecurity and deploy advanced AI systems for national security applications.
The allegations come as China accelerates efforts to match the capabilities of leading US AI models. Anthropic recently restricted access to its powerful Mythos model outside select markets, prompting concern among Chinese technology leaders.
At a cybersecurity conference in Beijing, 360 Security Technology founder Zhou Hongyi reportedly described Mythos as a “cyber nuclear weapon” and warned that China cannot afford to lag in frontier AI development. He called for the country to develop its own equivalent models, arguing that relying on foreign AI systems would create a strategic disadvantage.
The accusations mark the latest flashpoint in the growing AI rivalry between the United States and China, with frontier AI models increasingly becoming a focus of geopolitical competition as well as commercial innovation.
Anthropic is urging US lawmakers to introduce stronger protections against AI model distillation following the alleged campaign.(Image: AI-generated) 
Share
