Het verhaal
On 10 Sep 2026 Anthropic published “Detecting and countering misuse of AI: September 2026,” its most detailed threat-intelligence report to date, covering disrupted misuse of Claude from December 2025 through August 2026. Seven harm areas: cyber operations, influence operations, surveillance, scams/fraud, biological misuse, conventional weapons development, and model distillation. Actors included suspected state-sponsored groups, financially motivated criminals, spyware vendors, state propaganda outlets, and individuals. Standout cyber cases: GTG-20006 (Russian-nexus espionage with AI-rebuilt malware and hotel Wi‑Fi DNS hijacks), ShinyHunters-linked smash-and-grab affiliates, Chinese-speaking GTG-10007 exploit foundries/agent swarms, and hacktivist GTG-50029 against European political targets. Influence ops spanned Russian FIMI in CAR, a France-based ~70-site commercial influence network, a Malaysia election platform, Iranian state-aligned soft-war planning, and more. Anthropic says it disrupted every operation in the report, hardened safeguards, and shared intel with authorities and industry where appropriate. Full PDF linked from the post.