The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Google DeepMind launched Gemini 4 Carbon, an advanced AI capable of autonomously identifying and exploiting software vulnerabilities. While achieving record performance in cybersecurity tests, Google is restricting access due to concerns that its dual-use capabilities could enable cyberattacks if misused. No actual harm has occurred yet.[AI generated]
Why's our monitor labelling this an incident or hazard?
The AI system (Gemini 4 Carbon) is explicitly described as capable of discovering and exploiting software vulnerabilities, which could lead to significant harms such as cyberattacks, breaches of security, and disruption of critical infrastructure. Although no actual harm has been reported yet, the potential for such harm is credible and recognized by Google, which is limiting access and implementing safeguards. This fits the definition of an AI Hazard, as the development and use of this AI system could plausibly lead to an AI Incident involving harm to property, communities, or critical infrastructure.[AI generated]