Anthropic Reports Up to 99% Reduction in AI Model Distillation Attacks on Claude
Anthropic has implemented new security measures that have reduced data extracted through AI model distillation attacks on its Claude models by 90 to 99 percent. These safeguards restrict access to Claude's reasoning capabilities and block repeated probing via proxy networks. The company highlighted risks from unauthorized copying, citing Chinese AI firms like Zhipu and Moonshot as examples. Anthropic emphasized that such distillation poses biosecurity, cybersecurity, and national security concerns, and it has banned thousands of accounts linked to these activities.
First-hand measurement across 3 sources
We measured how 3 outlets covered this story. No outlet gave this story a measurable political slant — there is no left–right reading to report. Overall sentiment is positive (64/100). Lens Score 46/100.
Outlets measured: firstpost, businessstandard, economictimes. See how each one headlined and framed the same story in the source comparison below.
AI Analysis
Sentiment was consistent across outlets (62–65/100), indicating broadly factual reporting rather than editorialising.
Coverage timeline
economictimes broke this story on 13 Oct, 02:50 am. Other outlets followed.