On Monday April 20, an outage left thousands of ChatGPT users locked out of their AI assistant as OpenAI scrambled to restore service across multiple regions. The disruption peaked with over 5,000 user reports on Downdetector and revealed stark regional differences in impact.
Users began experiencing problems around 10am ET Monday morning, with the monitoring site showing a rapid escalation of complaints. By early afternoon Pacific Time, more than 5,000 people had reported being unable to load both ChatGPT and its companion tool Codex through OpenAI's platform.
The outage displayed unusual geographic patterns according to Downdetector data. Reports from the United Kingdom peaked at over 8,000 incidents while the United States showed only about 1,875 during the same period. This fourfold difference suggests either regional infrastructure vulnerabilities or timezone-based usage patterns that concentrated the disruption's visibility.
For affected users, the problems manifested differently depending on their activity when the outage struck. A TechRadar poll found that 63 percent of impacted users couldn't access their previous conversations while only 27 percent experienced login failures. Those already logged in faced issues with ongoing chats and image generation features while others found themselves completely locked out at authentication.
OpenAI's status page initially described the incident as "degraded performance" before upgrading it to a "partial outage" affecting nearly all components including conversations, login systems, voice mode functionality and image generation tools. The company confirmed engineers were investigating issues preventing access to ChatGPT and Codex across web browsers and mobile applications.
Indian users faced particularly severe access problems with data showing approximately 900 reports from that region alone. About 80 percent of Indian users struggled specifically with web browser access according to monitoring data. The United States saw nearly 1,900 reports while Canada recorded over 200 connectivity complaints during the disruption's peak around 8 p.m. local time.
This marks at least the second significant service interruption for OpenAI's flagship product this year following similar technical problems in February that caused high error rates and prevented new conversation initiation. That earlier incident also left users unable to start fresh chats or maintain existing ones reliably.
By late Monday afternoon Eastern Time, OpenAI announced it had deployed a fix and was "monitoring the recovery" of services. Downdetector reports gradually returned to baseline levels across both US and UK regions as engineers continued investigating residual issues affecting some components.
The rapid social media reaction highlighted how deeply integrated ChatGPT has become in daily workflows worldwide. "Is ChatGPT down?" became a trending search query on Google while X users oscillated between genuine panic about disrupted productivity and self-deprecating humor about their dependence on artificial intelligence assistance.
