Digitnomics 
By Admin,
Artificial intelligence company Anthropic has accused Chinese technology giant Alibaba of using nearly 25,000 fraudulent accounts to gain unauthorized access to its Claude AI model.
In a letter to the U.S. Congress, Anthropic’s Head of Policy, Sarah Heck, alleged that operators linked to Alibaba and its AI research arm, Qwen, generated more than 28.8 million interactions with Claude between April 22 and June 5, 2026.
According to the company, the activity violated its terms of service and was aimed at extracting Claude’s most advanced capabilities, including agentic reasoning, software engineering and long-horizon task execution.
Anthropic described the operation as a large-scale “distillation attack,” in which outputs from advanced AI models are harvested to improve competing systems without incurring the cost of developing them from scratch.
The company said the alleged campaign followed similar attacks it disclosed in February involving Chinese AI firms DeepSeek, Moonshot AI and MiniMax, which together generated more than 16 million Claude interactions through 24,000 fraudulent accounts.
Anthropic warned that such activities could undermine billions of dollars invested in AI research and development by U.S. companies.
It has urged the U.S. Congress to strengthen information sharing among AI developers, tighten restrictions on China’s access to advanced AI chips and impose penalties on organisations involved in AI model distillation attacks.
OpenAI has also previously accused Chinese AI laboratories of attempting to improperly access and replicate the capabilities of leading U.S. AI models.
The allegations are likely to intensify the ongoing AI rivalry between the United States and China as competition for leadership in frontier artificial intelligence continues to grow.
