Anthropic之恶
2026-09-18
m.21jingji.com
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Anthropic and OpenAI have accused several Chinese AI companies of illicitly extracting outputs and internal reasoning chains from their advanced language models, such as Claude, using unauthorized means like stolen credentials and fake accounts. This large-scale distillation has led to intellectual property and privacy violations, prompting U.S. agencies to issue warnings and companies to strengthen defenses.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves the use and misuse of AI systems (large language models) where unauthorized actors extract model outputs and internal reasoning chains to train their own models, constituting a violation of intellectual property rights and privacy. The article details actual occurrences of such unauthorized distillation, the detection and blocking measures taken by AI companies, and the resulting harms. Therefore, this qualifies as an AI Incident because the AI system's use and misuse have directly led to violations of intellectual property rights and potential privacy harms.[AI generated]