ChatGPT experiences outage – but services starting to return

ChatGPT, the widely used artificial intelligence chatbot developed by OpenAI, suffered a significant outage on June 10, disrupting services for users worldwide.

The incident, which began around 7:45 AM BST, led to widespread reports of elevated error rates, high latency, and complete service failures across the platform, affecting both ChatGPT’s free and paid tiers, as well as OpenAI’s APIs and its image generator, Sora.

Reports on outage tracking site DownDetector surged, peaking at nearly 2,000 around 11:00 AM ET (4:00 PM BST) as users in Europe, North America, and Australia came online and encountered issues. Users reported everything from being unable to load the service to encountering error messages, disrupting workflows and communication for many.

OpenAI quickly acknowledged the problem, confirming it was under investigation. In a statement on X at 10:22 AM ET (3:22 PM BST), the company stated: “Our engineers have identified the root cause and are working as fast as possible to fix the issue.”

By the afternoon, signs of recovery began to emerge. DownDetector reported a significant decrease in active reports, dropping to just over 500 before 2:30 PM ET (7:30 PM BST), indicating a partial restoration of services. OpenAI’s website also confirmed that it is “monitoring the fix” and “working towards full recovery,” though it did not specify a full return-to-service timeline.

While some users may still experience intermittent issues as the fix propagates, the rapid response from OpenAI and the declining number of reports suggest a gradual return to normalcy for the popular AI services after several hours of global disruption.