Distillation Attacks: Anthropic vs. Alibaba
Anthropic has accused Alibaba of conducting a massive operation to illicitly extract capabilities from its Claude artificial intelligence model, calling it the largest known distillation attack against the company to date, according to a letter seen by Reuters.
In the letter, sent to U.S. senators and White House officials, Anthropic alleged that operators linked to Alibaba's Qwen AI lab used nearly 25,000 fraudulent accounts to generate 28.8 million exchanges with Claude between April and June, bypassing regional restrictions and terms of service.
A Familiar Playbook at Greater Scale
The accusation follows a pattern Anthropic first documented publicly in February, when it accused three other Chinese AI labs — DeepSeek, Moonshot AI, and MiniMax — of running "industrial-scale campaigns" using 24,000 fraudulent accounts and 16 million queries to distill Claude's capabilities into their own models. The Alibaba-linked operation appears to dwarf those earlier efforts in both volume and sophistication.
Distillation involves sending large numbers of carefully crafted prompts to extract a model's reasoning patterns and outputs, which can then be used to train competing systems at a fraction of the original development cost. Anthropic said the campaign targeted Claude's most valuable features, including software engineering and agentic reasoning capabilities.
Escalating Tensions Over AI Security
The accusation lands amid heightened friction between Anthropic and the Trump administration over AI security. Earlier this month, the Commerce Department imposed export controls requiring Anthropic to restrict access to its newest models, Fable 5 and Mythos 5, to U.S. citizens only — a move linked in part to concerns about Chinese access to advanced AI systems. Anthropic responded by pulling both models from the market entirely.
Anthropic does not offer commercial access to Claude in China, making the alleged use of fraudulent accounts and proxy services necessary for any large-scale extraction effort. The company has previously warned that such distillation poses national security risks, since extracted capabilities could fuel offensive cyber operations.
Market and Policy Implications
Alibaba shares fell roughly 3% following reports of the accusation. The disclosure is likely to intensify Washington's scrutiny of Chinese access to American AI systems, a concern that has already prompted export controls and bipartisan congressional attention in recent months. Anthropic characterized the Alibaba operation as the most substantial effort by a Chinese entity to exploit advances made by leading U.S. research labs.
