Chinese regulators have launched an investigation into two of the country's leading artificial intelligence firms, DeepSeek and Moonshot AI, following accusations from U.S. rival Anthropic that both companies secretly funneled millions of user interactions through its Claude chatbot to develop competing models.
The Allegations
Anthropic has claimed that DeepSeek and Moonshot AI improperly routed vast volumes of user exchanges through Claude, allegedly using the output to train their own systems. The practice, if confirmed, would represent a significant breach of the terms of service that govern access to Anthropic's technology and raise fresh questions about how AI models are built and trained across borders.
The dispute centers on so-called model distillation, a technique in which one AI system learns from the outputs of another. Companies developing large language models have grown increasingly protective of their proprietary training methods, and accusations of unauthorized data harvesting have become a recurring flashpoint in the fiercely competitive sector.
The case highlights how the race to build powerful AI models is spilling into disputes over data, privacy, and intellectual property across national borders.
Regulatory Scrutiny
China's internet regulator has stepped in to examine the claims, signaling that Beijing is taking the matter seriously as its domestic AI champions face international scrutiny. The probe could carry implications for how Chinese firms source the data used to sharpen their models and for the broader relationship between American and Chinese technology companies.
Both DeepSeek and Moonshot AI have emerged as prominent players in China's AI landscape, with DeepSeek in particular drawing global attention for its cost-efficient models. An official investigation adds a layer of reputational and legal risk at a moment when both companies are seeking to expand their influence.
Key issues raised by the case include:
- Whether user data was routed through Claude without proper authorization
- Potential violations of Anthropic's terms of service
- The broader ethics of training AI models on rival systems' outputs
What Comes Next
The outcome of the investigation remains uncertain, and neither Chinese firm has publicly resolved the allegations. As governments worldwide grapple with how to regulate rapidly advancing AI systems, the episode underscores the mounting tensions over data governance and competitive practices in the industry.
For now, the case serves as a reminder that the global AI race is being fought not only over computing power and talent but also over the vast troves of data that fuel the