On September 3, a significant number of users across the globe experienced technical disruptions while attempting to access OpenAI’s popular AI tools, specifically ChatGPT and Codex. Data from Downdetector highlighted the scale of the incident, noting that 73 percent of user reports were centered on ChatGPT, while an additional 15 percent struggled with Codex. The outage appeared to be widespread, with notable reports of service failures coming from the UAE, including major hubs like Dubai and Abu Dhabi.

OpenAI acknowledged the instability shortly before 7:00 PM GST, confirming they were actively investigating the underlying causes. By 7:43 PM GST, the company provided a status update, clarifying that they had identified the problem, which was manifesting as degraded performance across their platforms. Since then, developers and casual users alike have been waiting for full stability as the team works to monitor and finalize the recovery of these essential AI services, which include everything from conversational models to image generation tools like DALL-E.