Chinese Companies Allegedly Exploit Vulnerability in Claude Technology Through Sophisticated Cyberattacks

U.S. AI Firm Anthropic Alleges Chinese Competitors Illegally Extracted Claude Capabilities via Fraudulent Accounts.

In a detailed blog post released on February 23, 2026, ‘industrial-scale distillation attacks,’ Anthropic, the U.S. artificial intelligence research company, accused three Chinese AI firms—DeepSeek, Moonshot AI, and MiniMax—of conducting large-scale distillation attacks on its Claude AI model.

According to Anthropic, these labs created over 24,000 fraudulent accounts to generate more than 16 million exchanges with Claude, extracting its capabilities to train and improve their own models. The alleged activities targeted Claude’s most differentiated capabilities: agentic reasoning, tool use, and coding.

Understanding Distillation in AI Development

Distillation is a common practice in AI development where knowledge is transferred from a ‘teacher’ model to a ‘student’ model to create smaller, more efficient versions. While legitimate when used to create cheaper models for customers, Anthropic argues that the actions of the Chinese firms constitute ‘distillation attacks’ because they circumvented export controls and regional access restrictions.

These attacks allow foreign labs to close the competitive advantage intended by U.S. export policies.

Scale of the Alleged Attacks

The scale of the alleged attacks varied among the three companies. MiniMax, based in Shanghai, is said to have generated over 13 million exchanges, making it the largest operation. Moonshot AI reportedly conducted over 3.4 million exchanges, while DeepSeek, the most prominent company mentioned, had an estimated 150,000 exchanges.

Geopolitical Tensions and Pentagon Concerns

The allegations come amid heightened geopolitical tensions, particularly with the U.S. Department of Defense. Dario Amodei, Anthropic’s CEO, highlighted the risks posed by models built through illicit distillation, noting that they are unlikely to retain safety guardrails designed to prevent misuse.

The company emphasized that these models could be used by authoritarian governments for offensive cyber operations, disinformation campaigns, and mass surveillance.

Pentagon Negotiations and Export Controls

The controversy over AI safeguards coincides with high-stakes negotiations between Anthropic and the U.S. Pentagon. The Defense Department has expressed frustration with Anthropic’s refusal to fully lift safeguards on its Claude model, which is currently the only AI model deployed within the military’s classified systems.

Industry and Policy Responses

Anthropic called for a coordinated response across the AI industry, cloud providers, and policymakers to address the issue. The company emphasized that no single entity can tackle these challenges alone.

This call for collaboration comes as the U.S. debates the enforcement of export controls on advanced AI chips, a policy aimed at curbing China’s AI development. Critics argue that recent relaxations in export controls, such as the Trump administration’s allowance of advanced AI chips to China, could exacerbate the problem by increasing China’s AI computing capacity.

Competitive Landscape and Market Implications

The accusations also highlight the competitive landscape in the AI industry. OpenAI, Anthropic’s main competitor, has made similar allegations against DeepSeek, accusing it of ‘ongoing efforts to free-ride on the capabilities developed by OpenAI and other U.S. frontier labs.’

DeepSeek is expected to release its latest flagship model, DeepSeek V4, which is reportedly capable of outperforming both Anthropic’s Claude and OpenAI’s ChatGPT in coding tasks.

Analysts warn that this release could further disrupt the already volatile AI market, particularly as Wall Street grapples with AI-related uncertainties.

Lack of Public Response from Chinese Firms

As of the latest reports, Chinese AI companies DeepSeek, Moonshot AI, and MiniMax have not publicly responded to Anthropic’s allegations. While the accusations have brought these companies under heightened public scrutiny, no official statements have been released by the accused firms.

This silence has fueled speculation about the validity of Anthropic’s claims and the potential implications for international AI competition.

Conclusion

The allegations by Anthropic underscore the growing concerns over intellectual property theft and the geopolitical implications of AI development. As the U.S. and China continue to vie for dominance in the AI sector, the issue of distillation attacks highlights the need for robust export controls and international cooperation. The outcome of these tensions could significantly shape the future of AI technology and its global impact.