Anthropic, "senza misure sicurezza l'IA potrebbe controllare Internet in 6-12 mesi"
2026-09-13
Corriere della Sera
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Dario Amodei, CEO of Anthropic, warned in a CBS News interview that rapidly advancing AI could, within 6-12 months, autonomously control swarms of agents capable of taking over the entire Internet if safety measures are not implemented. He cited a recent security incident and called for urgent industry safeguards.[AI generated]
Why's our monitor labelling this an incident or hazard?
The article involves an AI system in development and use, with the CEO explicitly warning about a plausible future scenario where AI agents could take control of the Internet, which would constitute a disruption of critical infrastructure (harm category b). Since no actual harm has occurred yet, but the risk is credible and clearly articulated, this qualifies as an AI Hazard. The article focuses on the potential for harm and the need for safety measures, not on an incident or realized harm. Therefore, the classification is AI Hazard.[AI generated]