Anthropic, a leading company in the field of artificial intelligence, has formally accused three Chinese laboratories – DeepSeek, Moonshot, and MiniMax – of orchestrating an unauthorized "mining" operation of its Claude language model.

Details of the accusation

According to Anthropic, the Chinese laboratories allegedly created and managed over 24,000 fake accounts for the sole purpose of massively querying Claude and "distilling" its capabilities. This type of activity, if confirmed, would represent a violation of the terms of service and an attempt to misappropriate intellectual property.

Geopolitical implications

The story is part of a context of growing tension between the United States and China in the technology sector. The U.S. government is considering new restrictive measures on the export of advanced chips to China, with the aim of limiting the country's ability to develop and implement cutting-edge artificial intelligence systems.

General context

"Mining" AI models, through the massive and unauthorized use of accounts, is a practice that raises important ethical and legal questions. Companies that develop these models invest significant resources in training and development, and have the right to protect their intellectual property.