Anthropic accuses Alibaba of a 'distillation attack,' claiming Claude received over 28.8 million accesses.
It has been revealed that Anthropic, the AI company behind the 'Claude' series, has sent a letter of accusation to White House officials and senators alleging that it is being subjected to a 'distillation attack' by Qwen Lab, the AI research institute of Chinese big tech company Alibaba.
Anthropic says Alibaba illicitly extracted Claude AI model capabilities | Reuters
Anthropic accuses Alibaba of campaign to extract AI capabilities
https://www.cnbc.com/2026/06/24/anthropic-alibaba-distillation-campaign.html
Anthropic accuses Alibaba of running largest distillation campaign against Claude
https://thenextweb.com/news/anthropic-accuses-alibaba-distillation-claude-qwen
'Distillation' is a technique that extracts the input and output of high-performance AI models and uses them to improve the performance of other models. Anthropic is working with OpenAI and Google to counter 'adversarial distillation attacks' by Chinese companies.
OpenAI, Google, and Anthropic are collaborating to counter 'hostile distillation attacks by Chinese AI companies' - GIGAZINE
According to Anthropic, distillation attacks carried out by small AI ventures such as DeepSeek, MiniMax, and Moonshot AI before February 2026 involved 24,000 fake accounts and a total of 16 million transactions.
However, the distillation attack by Qwen lab between April and June 2026, as pointed out by Anthropic, involved more than 28.8 million interactions with Claude using approximately 25,000 fake accounts, surpassing previous attacks.
An Anthropic spokesperson said, 'We believe that government and industry collaboration is necessary to combat illegal distillation, and we will continue to work with Congress and the executive branch to ensure that the United States maintains its leadership in the field of AI.'
Alibaba's spokesperson declined to comment.
In June 2026, Alibaba released 'Qwen3.7-Plus,' which boasts performance comparable to Anthropic's 'Claude Opus 4.6.'
Related Posts:
in AI, Posted by logc_nt