Image: Marco Verch // CC2.0 License

Microsoft’s Bing Search engine and several associated services, including ChatGPT’s web version, experienced a significant outage on Thursday, May 23, wherein a technical malfunction caused a cascading outage that impacted several key Microsoft services. With Bing, the company’s flagship search engine — down — numerous users worldwide struggled to access information online.

The disruption extended beyond search functionalities, affecting Microsoft Copilot, the GPT-4 powered AI chatbot, and the DALL-E based Bing Image Creator tool. According to media reports, the outage began around 3:00 AM Eastern Time, primarily impacting users in Asia and Europe. Reports on Downdetector.com, a platform that tracks online service outages, indicated a surge in user complaints starting at that time. Users attempting to access Copilot are encountering loading issues followed by error messages stating, “We are sorry but we’re unable to connect to the service at this time.”

The problems began around 3AM ET (12:30 PM UTC), causing disruptions that affected users globally, particularly in Asia and Europe. While some US users have reported intermittent restoration of services, many in other regions continue to experience difficulties. The outage persisted for several hours, with partial functionality beginning to return around 7:20AM PT.

Users visiting Bing.com are met with various errors. Some see a grey-colored webpage with the Bing search pane, while others receive messages indicating that “Bing isn’t available right now, but everything should be back to normal very soon.” The outage also rendered Microsoft Copilot, the company’s AI-powered development assistant, inoperable. Users trying to utilize Copilot for coding tasks were met with a frustrating “Loading” screen followed by an error message. Similarly, the Bing Image Creator tool, known for its AI-powered image generation capabilities, displayed error messages, mirroring the issues with Bing Search.

And if this was not enough, owing to Bing’s Application Programming Interface (API), which allows other platforms to integrate Bing’s search functionalities, the disruption cascaded to other search engines and AI tools as well. DuckDuckGo, a popular privacy-focused search engine that relies on Bing’s API for search results, encountered difficulties displaying search results. Users attempting to search on DuckDuckGo were greeted with error messages. The outage also affected the internet search functionality offered to Plus subscribers of ChatGPT, the prominent AI chatbot created by OpenAI which is all but a household name today. “Between around 10:10 PM PT yesterday and 6:50 AM PT today, we experienced a partial outage affecting ChatGPT’s web-browsing capabilities due to Bing being unavailable,” OpenAI noted.

The outage has prompted a wave of reactions on social media, with many users turning to platforms like Twitter to voice their frustrations. The hashtag #BingDown trended for several hours as users shared memes and complaints about the disruption. Reports on Downdetector showed that over 2,000 users in the US alone were affected, and that 56% of users had issues with logging in. 34% of them had troubles with searching online, while the remaining 10% had issues with logging in.

For its part, Microsoft acknowledged the issues with Copilot in a thread on X, stating that they are “working to isolate the cause of the issue.” However, the company has yet to provide a comprehensive statement addressing the broader outage affecting Bing and its API. The Microsoft service health dashboard does not indicate any other service outages. “While we isolate the root cause, we’re transitioning requests to alternate service components to expedite service recovery. More information can be found in the admin center under CP795190,” the company noted, adding, “We’re broadening our mitigation efforts following signs of partial recovery for http://copilot.microsoft.com and other features. We’re closely monitoring to determine if further actions are needed to resolve the issue. Further details can be found in the admin center under CP795190.”