US AI firm Anthropic claims Alibaba extracted Claude AI capabilities through millions of interactions, escalating tensions in the global AI race.
The rivalry between China and the US has not been hidden. Now, the American tech giant Anthropic has accused China’s Alibaba of carrying out a large-scale operation to extract capabilities from its Claude AI model. As per Reuters, the Dario Amodei-led Anthropic has sent a letter to the US authorities and the startup claims that this activity took place between April 22 and June 5, 2026, using nearly 25,000 fake accounts that allegedly generated over nearly 28.8 million interactions with Claude.
Here’s What Happened
As per Reuters, Anthropic alleges that Alibaba’s effort was an attempt to improve its own AI system through a technique called ‘distillation’.
Those who are not aware, distillation is a process in which a smaller or lesser advanced AI model learns from the outputs of a more capable system. AI companies often use this technique internally, but Anthropic argues that unauthorised use of its models to train rival systems violates its policies.
Reportedly, the American tech startup also claimed that people linked to Alibaba and its AI research division, Qwen used fraudulent accounts to access Claude.
Notably, the letter was sent on June 10 to US Senators Tim Scott and Elizabeth Warren, who lead the Senate Banking Committee. These allegations surfaced ahead of a hearing focused on AI and national security.
To recall, Anthropic had also raised similar concerns earlier this year. As per Reuters, it had identified similar activities involving Chinese AI firms including MiniMax, DeepSeek and Moonshot AI. Moreover, it also alleged that those attempts ranged from hundreds of thousands to millions of interactions with Claude.
