Anthropic Accuses Alibaba of Using 25K Fake Accounts to Steal Claude AI Data
Anthropic accuses Alibaba used 25,000 fake accounts to scrape Claude AI data, raising serious concerns over AI security and ethics.
Jersey City, N.J., June 25, 2026: Anthropic accused Chinese tech giant Alibaba this week of using nearly 25,000 fake accounts to collect data from its Claude artificial intelligence (AI) chatbot as part of an effort to improve Alibaba’s own AI models, escalating a dispute between U.S. and Chinese AI developers. Anthropic said the campaign was conducted between April 22 and June 5, 2026, and generated more than 28.8 million exchanges with Claude through accounts allegedly linked to Alibaba’s Qwen AI research unit, according to a Reuters report.
The allegations were outlined in a June 10 letter sent by Anthropic to U.S. Senators Tim Scott and Elizabeth Warren, which became public this week ahead of a Senate hearing on AI competition. The company said the activity was intended to support a technique known as AI distillation, in which responses from a more advanced AI system are used to help train or improve another model, according to an Anthropic post. The allegations, which Alibaba has not yet responded to publicly, remain unverified by independent parties.
According to Anthropic, the campaign focused on collecting responses related to coding and complex problem-solving tasks. The company told lawmakers that large-scale collection of AI-generated outputs could help competitors improve their own systems without making the same level of investment in computing infrastructure, research, and development.
Reuters reported that Anthropic linked the activity to Alibaba’s Qwen AI division, one of China’s leading AI model developers. The company described the operation as the largest known attempt to extract capabilities from Claude through unauthorized access.
“We believe combating the threat of illicit distillation requires coordinated action between government and industry,” an Anthropic spokesperson said in a statement. The company said it would continue working with Congress and the administration on AI policy and security issues, according to CNBC.
The allegations come as AI companies increasingly focus on protecting their models and the knowledge embedded within them. According to CNBC, Anthropic has previously raised concerns about similar activities involving other Chinese AI developers, reflecting broader industry concerns around AI model security and intellectual property.
Alibaba had not publicly responded to the allegations when the reports were published. The claims have not been independently verified, and no court or regulator has concluded that Alibaba engaged in unlawful conduct. Additional reporting from BBC News noted that concerns around AI model copying and unauthorized use are becoming increasingly important as AI systems grow more capable and commercially valuable.
For business and technology leaders, the case highlights how AI competition is expanding beyond model performance to the protection of intellectual property and proprietary capabilities. As companies invest billions of dollars in developing advanced AI systems, protecting the knowledge embedded within those models is becoming increasingly important. The dispute also reflects growing concerns across the industry about how AI capabilities can be replicated, shared, and protected as competition intensifies.
