Anthropic, a leading U.S. Artificial intelligence firm, has publicly accused three Chinese AI laboratories – DeepSeek, Moonshot AI, and MiniMax – of orchestrating large-scale campaigns to extract capabilities from its Claude chatbot, raising fresh concerns about intellectual property theft and the competitive landscape in the rapidly evolving AI industry.
The allegations, detailed in a blog post on Monday, center around the claim that the Chinese companies created approximately 24,000 fraudulent accounts to generate over 16 million interactions with Claude, a process known as “distillation.” Anthropic asserts that this activity violated its terms of service and regional access restrictions, as Claude is not commercially available in China.
Distillation is a common AI training technique where a smaller model learns to mimic the behavior of a larger, more sophisticated one. Even as Anthropic acknowledges distillation as a legitimate practice, it contends that the scale and method employed by DeepSeek, Moonshot, and MiniMax were specifically designed to illicitly acquire Claude’s proprietary knowledge. The company alleges the campaigns targeted areas where Claude excels, including complex reasoning, coding assistance, and tool use.
“These campaigns are growing in intensity and sophistication,” Anthropic wrote, adding that the situation demands “rapid, coordinated action among industry players, policymakers, and the global AI community.” The company has reportedly cut off known access points used by the Chinese firms and is advocating for tighter U.S. Export controls on advanced chips and AI services to prevent future occurrences.
The accusations follow similar claims made earlier this month by OpenAI, Anthropic’s rival, which alleged that DeepSeek and other Chinese AI companies were illegally distilling its ChatGPT models. OpenAI shared its findings with the U.S. House Select Committee on China, highlighting the potential national security implications of these practices. According to CNN reporting, DeepSeek’s rapid advancements, particularly its launch of a powerful model requiring fewer computing resources than competitors, had already raised questions about the effectiveness of existing U.S. Tech and export controls.
Anthropic’s claims come as the company itself recently settled a copyright lawsuit with thousands of authors for $1.5 billion in September 2025. The lawsuit alleged that Anthropic had downloaded books in bulk from shadow libraries to train its AI models, rather than acquiring them through legal channels. The settlement involved payments of approximately $3,000 per book for roughly 500,000 works.
The Chinese firms allegedly circumvented geofencing and business restrictions by routing traffic through proxy services that resell access to Western AI models. Anthropic described one such operation as a “hydra cluster” capable of simultaneously managing tens of thousands of accounts and distributing requests across multiple API keys and cloud providers. The labs then reportedly scripted lengthy conversations with Claude, designed to elicit detailed responses that could be used as training data for their own models.
Reaction to Anthropic’s accusations has been mixed, with some observers pointing to the company’s own past data collection practices. Commentators on social media noted the irony of Anthropic accusing others of questionable data acquisition methods, referencing the Reddit scraping case and the author copyright lawsuit. The broader debate centers on the ethical boundaries of data collection and the competitive dynamics in the AI industry, where firms are constantly seeking to improve their models by leveraging available data.
As of Tuesday, February 24, 2026, DeepSeek, Moonshot AI, and MiniMax have not publicly responded to Anthropic’s allegations. The U.S. Government has not yet issued a statement regarding the matter, but the accusations are likely to intensify scrutiny of China’s AI development efforts and fuel calls for stricter export controls.
