For the third time in two months, OpenAI’s ChatGPT has been down owing to a technical outage. The service disruption let millions of users across the globe unable to access the popular AI chatbot.
The service disruption began on Thursday, and went on the affect OpenAI’s associated services as well. This means that apart from individual users (the chabot currently has over 300 million weekly active users globally), businesses and organizations that are reliant on OpenAI’s API faced disruptions in their services as well. Many of these businesses rely on the API for applications such as customer service, content creation, and data processing, to name some.
The outage, which started around 5 PM IST, was immediately noticeable as thousands of users reported being unable to log into the chatbot. Downdetector, the popular outage-tracking website, recorded a sharp spike in complaints, with over 4,000 reports submitted during the peak of the disruption. Many users encountered a “Bad Gateway” error when attempting to access ChatGPT’s web platform, while others experienced degraded performance instead. As is the norm, they resorted to social media (mostly X) to air their frustration over the development.
Although OpenAI has not confirmed the exact reason behind the outage, media reports speculate that the outage could be linked to server overload, or even issues with the deployment of the company’s newer AI models, such as GPT-4o and GPT-4o mini. Take this with a grain of salt, since these speculations are yet to be officially confirmed by OpenAI.
OpenAI has responded to the development, though, acknowledging the outage on its official Status page. The company initially identified the issue as “degraded performance” and stated that it was investigating elevated error rates for both ChatGPT and its API.
“This issue has now been resolved. Between 4:23am and 7:10am PST, customers experienced elevated errors on ChatGPT,” read OpenAI’s update on the Status Page.
Eventual updates provided throughout the day indicated that the AI firm had identified the root cause and was actively working on a fix. By 10:09 AM ET, OpenAI reported that a resolution had been implemented and the service was being monitored.
OpenAI later confirmed that the problem was resolved between 3:33 AM and 4:23 AM PST, during which customers experienced intermittent errors. However, some users continued to report access issues even after the fix was deployed
There’s a silver lining, though – while the web version of ChatGPT was affected by the service disruption, the mobile application appeared to function normally. Reports indicate that the app was still able to process prompts and generate responses without any issues.
