Anthropic: Chinese firms extracted Claude data

 Anthropic PBC accused in a blog post on Monday some Chinese companies of "illicitly extracting" data from its Claude model to improve their own models.

DeepSeek, Moonshot, and MiniMax's AI labs "generated over 16 million exchanges with Claude through approximately 24,000 fraudulent accounts, in violation of our terms of service and regional access restrictions," Anthropic said, adding that they used "distillation," a technique that involves training a less capable model on the outputs of a stronger one. The business stressed that everything stolen in such operations can be sent to military, intelligence, and surveillance systems of foreign countries, "enabling authoritarian governments to deploy frontier AI for offensive cyber operations, disinformation campaigns, and mass surveillance."

Anthropic warned that such campaigns are getting more intense and more sophisticated, underlining that industry players, policymakers, and the global AI community need to act fast and in unison.

https://breakingthenews.net/Article/Anthropic:-Chinese-firms-extracted-Claude-data/65727366