Anthropic accuses DeepSeek and other Chinese firms of using Claude to train their AI

Source: The Verge

Background

Anthropic alleges that three Chinese AI companies misused its Claude model in an effort to improve their own products. The claim was announced on Monday, citing “industrial‑scale campaigns” that involved the creation of roughly 24,000 fraudulent accounts and more than 16 million interactions with Claude, as reported by The Wall Street Journal.

Accusations

Anthropic says the companies engaged in “distilling” Claude—training a smaller AI model based on a more advanced one. While distillation can be a legitimate training method, Anthropic warns that it can also be employed for illicit purposes.

Companies Involved

• DeepSeek
• MiniMax
• Moonshot

Further Reading

Read the full story at The Verge