Anthropic has asked US lawmakers to act against Alibaba. The company claims Alibaba ran a large AI distillation campaign against its Claude model.
The allegation centers on nearly 25,000 fake accounts. Anthropic says those accounts created about 28.8 million exchanges with Claude in six weeks.
In simple terms, AI distillation copies behavior from a stronger model. A smaller model learns from the larger model’s answers. It can then mimic key skills with lower computing costs.
Anthropic says the campaign did not steal source code or model weights. Instead, it allegedly harvested answers at massive scale. The company argues that this still extracts valuable model capability.
The reported activity targeted software development, multi-step reasoning, and agent-style tasks. These tasks let AI plan and complete complex workflows. That matters for telecom, UC, and VoIP teams using automation.
For communications providers, AI now supports routing, fraud detection, support, and security. If model behavior becomes easy to copy, vendors face new risk. Innovation may become harder to protect.
At the same time, distillation is not always harmful. Developers use it to make AI cheaper and faster. Smaller models can run in more places, including edge environments.
The dispute therefore focuses on permission, scale, and intent. Anthropic says Alibaba crossed a clear line. It wants US policymakers to create penalties for similar activity.
Francesco Pili, Counsel at Hogan Lovells, said the issue reaches beyond one company.
“This is part of a broader trend of regulatory disputes over whether, and on what terms, AI systems can access third-party platforms, services, or data,” he said.
“These cases are blurring the line between traditional platforms and agentic AI services. They are also testing the limits of what target platforms may lawfully do in response, including whether, and how, they can restrict access.”
The case also highlights growing tension between US and Chinese AI firms. Anthropic called unapproved extraction a “massive subsidy for our geopolitical competitors.”
The timing is important. Chinese AI developers have advanced quickly in recent years. DeepSeek previously surprised the market with powerful, lower-cost model performance.
Anthropic has made similar claims before. It previously accused DeepSeek, Moonshot AI, and MiniMax of generating millions of Claude exchanges. The alleged Alibaba campaign appears larger and more organized.
For enterprise technology leaders, the message is clear. AI access controls now need the same attention as network security. API monitoring, account verification, and traffic analysis will become essential.
This dispute may shape future AI policy and vendor contracts. It may also influence how communication platforms expose AI features. For now, Anthropic wants lawmakers to define firmer boundaries.
