The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
On July 19, 2026, ChatGPT by OpenAI experienced a global outage, affecting users in Ukraine, the US, Europe, India, and Japan. Over 250 complaints were reported, mainly from iOS users, citing unresponsiveness, slow processing, and errors like 502 Bad Gateway and 503 Service Unavailable.[AI generated]
Why's our monitor labelling this an incident or hazard?
ChatGPT is a generative AI system. The reported global outage and malfunction of the AI system have directly caused disruption to the service, affecting a broad user base and critical digital workflows. This fits the definition of an AI Incident as the AI system's malfunction has directly led to harm in the form of disruption to critical infrastructure and harm to communities relying on the service. The event is not merely a potential hazard or complementary information but a realized incident of AI system failure causing harm.[AI generated]