In brief: Anthropic accuses Alibaba of using Claude outputs to train its own models and asks the US government for support against such terms-of-service violations.
Anthropic has accused Alibaba of using Claude models to generate training data for its own AI systems without complying with the terms of service. The company is calling on the US government to take action against such violations.
Anthropic has made public that Chinese technology conglomerate Alibaba used its Claude models in a manner that violates the terms of service. Specifically, Alibaba is accused of using outputs from Claude systems as training data for its own AI models – a procedure known in machine learning as distillation.
This practice poses a substantial business risk for AI manufacturers. Through systematic distillation, features and behaviors of proprietary models can be transferred into cheaper or faster successor models without the original developers benefiting or retaining control over application. For Anthropic, this also means a loss of control over its own security and compliance investments.
Anthropic has therefore called on the US government to create a stronger political and regulatory framework to prevent such violations of terms of service. With this move, the company signals that legal and economic enforcement mechanisms at the bilateral level alone may not be sufficient – particularly given that Alibaba is a company headquartered in China, where US laws are difficult to enforce.
Source: www.golem.de · Published 25 June 2026
Lumi AI News — AI-assisted curation in accordance with Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.1.