Full Breakdown
Anthropic Accuses Alibaba of the Largest Known AI Distillation Attack
6/26/2026, 12:03:43 PM
The Alleged Distillation Campaign
Anthropic says operators linked to Alibaba and its Qwen AI lab used roughly 25,000 fraudulent accounts to query Claude — its flagship large-language model — 28.8 million times between 22 April and 5 June 2026. The campaign allegedly targeted Claude’s agentic reasoning, software-engineering, and long-horizon task capabilities, violating the model’s terms of service and access restrictions.
Background: AI Theft and U.S. Policy
In April 2026 the White House Office of Science and Technology Policy issued a memo labeling China’s “industrial-scale” AI theft as a national-security threat. Earlier in the year Anthropic disclosed three separate distillation campaigns by DeepSeek, Moonshot AI and MiniMax, totaling about 16 million illicit exchanges. The Alibaba effort, by Anthropic’s count, dwarfs those prior incidents.
Key Players
- Anthropic – U.S. AI firm behind Claude.
- Alibaba Group (9988.HK) – Chinese e-commerce and tech conglomerate; its AI division is Alibaba Qwen.
- Senators Tim Scott (R-SC) and Elizabeth Warren (D-MA) – Chair and ranking member of the Senate Banking Committee.
- U.S. Department of Defense & Pentagon – Designated Alibaba a “Chinese military company.”
- U.S. Commerce Department – Imposed export restrictions on Anthropic’s Mythos and Fable models.
Timeline of Events
- 22 Apr 2026 – Fraudulent accounts begin querying Claude.
- 10 Jun 2026 – Anthropic sends a confidential letter to Senators Scott and Warren.
- 12 Jun 2026 – Commerce Department restricts Anthropic’s Mythos and Fable models.
- 24 Jun 2026 – Reuters publishes Anthropic’s allegations; Alibaba added to Pentagon’s 1260H list.
- Late Jun 2026 – Alibaba files a lawsuit seeking removal from the Pentagon list.
Scale of the Operation
- ? 25 000 fraudulent user accounts.
- ? 28.8 million Claude interactions.
- Targeted capabilities: agentic reasoning, software engineering, long-horizon planning.
Why It Matters
Anthropic argues that illicit distillation lets rivals acquire frontier AI functions without the billions of dollars required for research, training, and compute. The alleged theft therefore converts U.S. R&D investment into a “massive subsidy” for geopolitical competitors and raises concerns about the export of advanced AI to foreign militaries. Market reaction saw Alibaba shares fall more than 4 percent in Hong Kong trading.
Official Statements & Responses
Anthropic’s spokesperson emphasized the need for coordinated government-industry action and pledged continued cooperation with Congress and the Administration. The company expressed support for the White House memo and for the Senate Banking Committee’s upcoming AI hearing. The Pentagon confirmed Alibaba’s inclusion on its Chinese military companies list, a designation Alibaba is contesting in court. The Commerce Department cited potential military-intelligence misuse as the reason for restricting Anthropic’s Mythos and Fable models, prompting Anthropic to disable global access.
Criticism & Opposition
Alibaba has not provided a comment to multiple news outlets, effectively denying the allegations without a public response. Analysts note that the Senate Banking Committee’s jurisdiction over foreign IP theft is limited, making enforcement of any congressional penalties uncertain. Some observers view the reputational impact on Alibaba as modest, given prior similar accusations against other Chinese AI firms.
Conflicting Reports & Gaps
Anthropic’s letter claims Alibaba “ignored the Trump Administration’s warnings,” yet the Trump administration later imposed export controls on Anthropic’s own models, creating a policy inconsistency. Additionally, while the Commerce Department barred DeepSeek from a trade blacklist, it has not taken comparable action against Alibaba, highlighting selective enforcement.
Verbatim Quotes
- “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,” — Anthropic spokesperson
- “Distillation attacks turn hundreds of billions of dollars in American investment and [research and development] into a massive subsidy for our geopolitical competitors,” — Anthropic
- “ignored the Trump Administration’s warnings” — Anthropic letter
- “These distillation attacks are carried out illicitly, systematically, and at industrial scale to harvest US AI capabilities across frontier labs and repackage them as their own.” — Anthropic
- “ Alibaba was added to the Pentagon's Chinese military companies list this month, a designationit is challenging.” — Reuters
What’s Next
The Senate Banking Committee will hold its AI hearing on 30 June, where lawmakers are expected to discuss legislative measures to curb illicit model distillation. Alibaba’s lawsuit to remove the Pentagon designation is pending, and further U.S. export controls on advanced AI hardware may follow.
