Anthropic Accuses DeepSeek and Other Chinese AI Models of Data Theft

Anthropic, the US based AI firm behind Claude, has accused three Chinese AI companies of illicitly extracting data from its chatbot. The allegations come amid growing tensions between US and Chinese AI developers. 

Anthropic's Allegations Against Chinese AI Firms 

On February 24, 2026, Anthropic claimed in a blog post that DeepSeek, MiniMax, and Moonshot AI created 24,000 fraudulent accounts to interact with Claude. These accounts reportedly generated over 16 million exchanges, extracting high quality outputs in coding, reasoning, and tool use. Anthropic described this as "industrial scale" data distillation, comparing it to copying answers during an exam. 

The company argued that while distillation is a legitimate technique when done internally, unauthorised use to replicate proprietary AI models violates its terms. Anthropic warned that models built this way might lack proper safeguards, potentially spreading dangerous capabilities. 

Public Backlash Against Anthropic 

Instead of garnering support, Anthropic faced criticism online. Social media users pointed out the company's $1.5 billion settlement in a lawsuit where authors accused it of using pirated books to train Claude. One X user noted Claude reproduced content verbatim from their Traditional Chinese Medicine website without permission. 

Elon Musk joined the discussion, tweeting that Anthropic had "stolen training data at massive scale" and paid multi billion dollar settlements. While acknowledging his own company xAI also trains on large datasets, Musk called Anthropic "hypocritical" for its stance. 

Read More: India’s Share of Global AI Funding Stands at 0.6% in 2025! 

Geopolitical Context of the Dispute 

The accusations coincide with heightened US China tensions in AI development. Reuters reported DeepSeek allegedly used restricted Nvidia Blackwell chips in Inner Mongolia, potentially bypassing US export controls. This comes as DeepSeek prepares to launch its R2 model, rumoured to rival top US AI systems. 

Chinese AI models pose a unique challenge as most are open source, allowing global access unlike proprietary US models. If Chinese models match Claude or ChatGPT in performance, it could impact the commercial value of American AI companies. 

Conclusion 

Anthropic's allegations highlight ongoing competition in AI development between US and Chinese firms. The public response underscores broader debates about data sourcing practices in the AI industry, with many viewing the situation as reflective of standard industry practices rather than exceptional misconduct. 

Disclaimer: This blog has been written exclusively for educational purposes. The securities or companies mentioned are only examples and not recommendations. This does not constitute a personal recommendation or investment advice. It does not aim to influence any individual or entity to make investment decisions. Recipients should conduct their own research and assessments to form an independent opinion about investment decisions. 

Investments in the securities market are subject to market risks, read all the related documents carefully before investing.