financePublished: 11 Sept 2026, 05:59Trust level: verified

Anthropic Says Chinese AI Labs Used Claude Exchanges to Train Models

Anthropic Says Chinese AI Labs Used Claude Exchanges to Train Models

Summary

Anthropic detected unauthorized use of millions of Claude interactions by China-based labs, including Alibaba and Moonshot AI, to train their models.

Prepared by the newsroom with AI assistance; the original link is always provided

AI company Anthropic announced that it detected unauthorized efforts by Chinese artificial intelligence labs to utilize its Claude models to enhance their own systems. The company stated that millions of exchanges were accessed to train and improve competing AI products.

According to Anthropic, the entities involved in these activities include major China-based organizations such as Alibaba and Moonshot AI. These labs allegedly used data from Claude interactions without permission to train their generative AI architectures.

Latest news