Anthropic has accused Alibaba and its affiliated Qwen AI lab of carrying out what it described as the largest known attack on its Claude system, alleging that the Chinese tech giant used a wide network of fraudulent accounts to probe and extract model capabilities.
In a letter seen by Reuters, Anthropic said the campaign ran from April 22 to June 5, 2026, and generated more than 28.8 million exchanges with Claude through nearly 25,000 fraudulent accounts. The company said the activity was aimed at lifting Chinese access to Anthropic’s advanced AI capabilities.
Also Read | Google to allow external payments on Play Store from June 30
🚨Anthropic Accuses Alibaba of Massive AI Model Extraction Campaign
— Bull Theory (@BullTheoryio) June 25, 2026
Anthropic has accused Alibaba of orchestrating the largest known attempt to extract capabilities from its Claude AI models through a massive network of fraudulent accounts.
The allegations were disclosed in a… pic.twitter.com/FLLccsaFcE
The science behind the squabble
Anthropic said the technique at the centre of the dispute is "distillation", a process in which a less capable model is trained on the outputs of a stronger one. In its account, the alleged campaign sought to copy capabilities associated with Claude’s more advanced functions, including software engineering and agentic reasoning, and to help accelerate access to Anthropic’s Mythos Preview capabilities.
Anthropic has previously said it uncovered similar efforts by other Chinese AI companies, including DeepSeek, Moonshot AI and MiniMax, which it said together accounted for more than 16 million exchanges and roughly 24,000 fraudulent accounts.
🚨 CHINA LE ROBÓ LA IA A ANTHROPIC.
— SONIA (@S0N_IA) June 24, 2026
Y lo hicieron durante 45 dĂas seguidos.
Anthropic acaba de acusar a Alibaba de lanzar el mayor ataque de robo de IA de la historia.
El objetivo: extraer las capacidades de Claude… sin pagar ni un dólar.
CĂłmo lo hicieron:
→ Crearon casi… https://t.co/KgB1aooXTC pic.twitter.com/ggisWbjFyw
A delicate dance of power and processors
The allegation lands at a sensitive moment in the US-China technology contest, where control over frontier AI systems and the data used to train them has become a growing policy issue.
Anthropic said in the letter that it supported US government efforts to counter such attacks, including threat-intelligence sharing with private-sector AI companies. The letter was sent on June 10 to Senators Tim Scott and Elizabeth Warren, who lead the Senate Banking Committee, before a scheduled hearing on artificial intelligence. Alibaba did not immediately respond to a request for comment.
Officialdom enters the fray
Reuters also noted that Anthropic’s complaint comes amid other policy and security pressures around the company and the broader sector.
It reported that Alibaba was added this month to the Pentagon’s Chinese military companies list, a designation the company is challenging, while the Commerce Department imposed restrictions on Anthropic’s latest Mythos and Fable models on June 12 over concerns that they could be used by military intelligence users in China and other countries of concern, forcing Anthropic to disable access globally.
Also Read | India climbs to 13th globally in QS Future Skills Index 2027, signalling a stronger AI-ready workforce
FAQs
Q1: What is AI distillation, and why is it controversial?
Ans: AI distillation is a technique in which a smaller model learns from a stronger model's outputs, and it becomes controversial when done without authorization.
Q2: Why has Anthropic accused Alibaba and Qwen AI?
Ans: Anthropic alleges they used thousands of fraudulent accounts and millions of interactions with Claude to extract its AI capabilities.
