AI BriefingAnthropic Accuses Alibaba of Trying to Steal Its Technology Through AI Model Distillation
Key point
Anthropic has accused Alibaba to the US Congress of carrying out a large-scale distillation attack that mobilized thousands of accounts to steal the performance of its AI models.
Details
AI startup Anthropic has sent an official letter to a U.S. Senate committee alleging that Chinese tech company Alibaba attempted to illegally extract its AI capabilities. Anthropic claimed that operators linked to Alibaba tried to steal its model's technology through a Distillation attack.
Distillation is an AI training method that extracts outputs from an existing powerful model to build a smaller, lower-performing model. According to Anthropic's announcement, operators affiliated with Alibaba used approximately 25,000 fake accounts to exchange 28.8 million conversations with its model between April 22 and June 5.
This incident occurred two months after the Trump administration called for cooperation to counter industrial-scale distillation attacks. Anthropic added that Alibaba carried out the attack despite the government's warning.
Meanwhile, Anthropic has recently received export control guidance from the Trump administration, on national security grounds, to restrict access to its latest models, Fable 5 and Mythos 5, and is currently in discussions with government officials to resolve the issue.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
US Appeals Court Upholds Pentagon Blacklisting of Anthropic as Supply Chain Risk