Anthropic Alleges Chinese AI Firms Copied Claude Data, Internet Pushes Back
Anthropic accuses Chinese AI firms of data theft, but online critics question its own training practices and legal history.
Tensions between American and Chinese artificial intelligence companies have taken a sharper turn after Anthropic accused three Chinese AI firms—DeepSeek, MiniMax, and Moonshot AI—of improperly extracting data from its Claude chatbot to improve their own models.
In a recent blog post, Anthropic claimed the companies orchestrated what it described as “industrial-scale” efforts to distill knowledge from Claude. According to the company, approximately 24,000 fraudulent accounts were created, generating over 16 million exchanges with its AI system. The alleged goal was to harvest high-quality responses in areas such as coding, reasoning, and tool usage—capabilities Anthropic considers core strengths of Claude.
In the AI industry, this technique is known as distillation—training a smaller or newer model using outputs from a more advanced one. While distillation is a common internal practice among AI developers, Anthropic argues it becomes unlawful when done without permission. “We have identified industrial-scale campaigns by three AI laboratories—DeepSeek, Moonshot, and MiniMax—to illicitly extract Claude’s capabilities to improve their own models,” the company stated.
Anthropic further warned that models developed through “illicit distillation” may not retain safety safeguards embedded in the original system. “Models built through illicit distillation are unlikely to retain those safeguards, meaning that dangerous capabilities can proliferate with many protections stripped out entirely,” the company said.
However, instead of rallying public sympathy, the accusations have sparked intense debate online. Many social media users pointed out that the broader AI industry—including major US players—has long relied on large-scale data scraping to train models.
Some critics referenced Anthropic’s own legal challenges, including settlements with authors who accused the company of training Claude on pirated books. “I have a website about Traditional Chinese Medicine that I spent literally years building. When I asked questions to Claude about the topic, it parroted almost word-for-word what I myself wrote. So please spare us the gaslighting about training AI on others' work,” one user wrote on X.
Another user commented, “This is pure evil. First, Anthropic used torrents to download books and data for training on top of the internet, without asking permission from anyone. Second, the Chinese companies paid for the API and did not get the information for free. Third, it seems like Anthropic is tracking people.”
Tech billionaire Elon Musk also weighed in, criticising Anthropic’s stance. “Anthropic is guilty of stealing training data at massive scale and has had to pay multi-billion dollar settlements for their theft. This is just a fact,” Musk wrote. While acknowledging that AI systems require massive datasets, he suggested Anthropic was being “smug” and “hypocritical.”
The controversy unfolds against a backdrop of escalating US–China competition in artificial intelligence. US officials increasingly frame advanced AI development as a national security issue. A recent Reuters report alleged that DeepSeek may have trained its latest model using advanced Nvidia Blackwell chips despite export restrictions, further intensifying scrutiny.
With DeepSeek reportedly preparing to launch its next-generation model—rumoured to rival top American systems—the dispute underscores not just a legal battle, but a high-stakes technological rivalry. As open-source Chinese models gain ground and proprietary US systems defend their commercial edge, the global AI race appears more competitive—and contentious—than ever.
