AI chatbot ChatGPT experienced a significant outage on Friday, leaving thousands of users, including those in Oxfordshire, unable to access the service. Reports indicated users encountered an "Unexpected status 401: Incorrect API key" error during the disruption.
ChatGPT Outage Details and User Reports
According to outage-tracking website Downdetector, reports of issues with OpenAI services began around 6:55 p.m. EDT on Friday. The Hindustan Times reported that users experienced an "Unexpected status 401: Incorrect API key" error while attempting to use ChatGPT and Codex.
Many users took to social media, with one reporting, "I was in the middle of a chat and I got the ‘Error in message stream’ with a RETRY button." Another noted, "ChatGpt is down. I am getting an error saying incorrect API key provided. I have never had this error before it literally just happened a few minutes ago."
OpenAI’s status page acknowledged the Codex error, stating they were "working on implementing a mitigation." The services appeared to be working normally again after a brief disruption, though OpenAI did not disclose the cause of this particular outage.
Broader Cloud Outage Raises Concentrated Risk Concerns
This incident follows a separate, more widespread cloud outage on September 3, which affected multiple prominent AI services including ChatGPT, Claude, Grok, and Microsoft’s own Copilot.
InfoWorld highlighted that this earlier outage was due to a failure in Microsoft Azure’s East US region. Downdetector logged over 37,000 reports for ChatGPT alone during that event, with combined reports climbing past 66,000.
The simultaneity of these failures, despite being from aggressively competing AI labs, revealed a "concentrated risk" due to shared regional dependencies on cloud infrastructure. Google’s Gemini, notably, remained largely operational during the September 3 incident because it runs on its own vertically integrated cloud, avoiding the common "attack surface."
Implications for Businesses and Future AI Adoption
This series of outages underscores concerns for businesses increasingly relying on AI. InfoWorld warned that while these were primarily consumer-facing chatbots, future failures of enterprise large language models and AI agents could have significant financial impacts.
When AI agents orchestrating finance workflows or supply chains go dark, the problem extends beyond a lost chat window, impacting automated decisions and transactions at scale. Enterprises are building dependency structures with AI and cloud services faster than they are mapping them, leading to potential unseen vulnerabilities. Even if an enterprise isn’t a direct customer of a major cloud provider, their core services often rely on vendors who are, creating a cascading failure risk.
For more on related developments, see our articles on OpenAI Pauses AI Model Work Amid Cyber Concerns and ChatGPT Down Globally, OpenAI Confirms Outage.
Frequently Asked Questions about Recent AI Outages
- Q: What was the main error users saw during the recent ChatGPT outage?
- A: Users frequently reported an "Unexpected status 401: Incorrect API key" error, along with messages like "Error in message stream."
- Q: What caused the Friday outage of ChatGPT and Codex?
- A: OpenAI did not disclose the specific cause of the outage on Friday evening, though services quickly returned to normal.
- Q: Were other AI services affected by recent outages?
- A: Yes, a separate incident on September 3 saw ChatGPT, Claude, Grok, and Microsoft Copilot affected due to a failure in Microsoft Azure’s East US region.
