Anthropic accuses Alibaba of using 25,000 fake accounts to steal Claude AI capabilities

Anthropic accuses Alibaba of using 25,000 fake accounts to scrape Claude AI

The alleged campaign generated 28.8 million queries in a large-scale model-extraction effort to replicate AI capabilities. Anthropic has accused Alibaba of using nearly 25,000 fraudulent accounts to extract capabilities from its Claude AI models, in what the US AI company described as the largest known attack of its kind against it. The campaign, carried out between April 22 and June 5, generated more than 28.8 million exchanges with Claude, according to a June 10 letter Anthropic sent to senior members of the US Senate Banking Committee, Reuters reported. Anthropic said the effort involved "distillation," a technique in which a less capable AI model is trained on the outputs of a more advanced system, potentially allowing rivals to replicate some of its capabilities at lower cost.

Anthropic accuses Alibaba of campaign to 'brazenly' and 'illicitly' rip off its AI capabilities

Anthropic accused the Chinese tech giant Alibaba of "brazenly" and "illicitly" trying to extract its artificial intelligence capabilities - carrying out "the largest known distillation attack on Anthropic to date." Anthropic detailed the alleged violation of its terms in a letter it sent to the Senate Committee on Banking, Housing and Urban Affairs earlier this month, Bloomberg reported. The act Anthropic imputed to Alibaba -- distillation -- is an AI training method in which a small, less capable model is built using outputs from an existing, more sophisticated model. Dario Amodei-led Anthropic wrote that operators affiliated with Alibaba and its Qwen AI lab conducted 28.8 million exchanges with its Claude models using about 25,000 fake accounts between April 22 and June 5, according to Bloomberg. "We believe combating the threat of illicit distillation requires coordinated action between government and industry, and we will continue working with Congress and the Administration to maintain American AI leadership," an Anthropic spokesperson said in a statement. Alibaba didn't immediately respond to The Post's request for comment. Anthropic's letter, addressed to Sens. Tim Scott (R-SC) and Elizabeth Warren (D-Mass.), came less than two months after the White House Office of Science and Technology Policy issued a memorandum pledging to assist AI companies in detecting and coordinating against mass distillation. Alibaba "ignored the Trump Administration's warnings," Anthropic claimed in its missive. A flood of powerful, cheap-to-use Chinese AI models have been winning over U.S. customers - and experts are warning that America's lead in the field could be in danger. One new open-source model, dubbed GLM-5.2, was released by China's z.AI on June 16 and specializes in coding projects. The company claims that GLM-5.2 is about as advanced as some of the best models offered by Anthropic, OpenAI and Google. Anthropic said in February that it had identified three "industrial-scale" distillation campaigns from three other Chinese AI labs: DeepSeek, Moonshot and MiniMax. Anthropic claimed the campaigns were becoming more sophisticated and intense. It also encouraged policymakers and other tech firms like cloud providers to collaborate to thwart future attacks. Recent weeks have seen Anthropic repeatedly clash with the White House, whoever. In the latest twist, the Trump administration launched a crackdown on the company's powerful "Fable" and "Mythos" models, prompting Anthropic to scramble to resolve security concerns.