OpenAI confirmed Tuesday that its ChatGPT service saw a spike in errors before it applied fixes. File Photo by Wu Hao/EPA
Dec. 2 (UPI) — ChatGPT malfunctioned for some users on Tuesday, before the company that operates the popular artificial intelligence chatbot announced fixes.
More than 3,000 users of ChatGPT reported problems getting the artificial intelligence model to work correctly, according to Downdetector. OpenAI confirmed in a post that it had seen increased error rates for its ChatGPT service on Tuesday afternoon Eastern time.
About a half hour after reporting the spike in errors, OpenAI posted an update stating that it had applied a “mitigation” and was watching to see how it played out. The cause of the outage and the exact number of users affected are not clear.
Some users reported getting blank responses from ChatGPT except for a blinking circle that typically indicates the service is processing a query, according to TechRadar.
Since its launch by OpenAI three years ago, ChatGPT has been at the center of the current AI boom with more than 800 million weekly users. The high-powered chatbot is built using vast amounts of text and produces human-like responses to queries on writing emails, cooking, research and other tasks.
The outage comes less than a week after OpenAI reported a limited security breach at Mixpanel, the company’s data analytics platform.
It also comes as ChatGPT faces growing competition from other tech giants. Last month, Google rolled out Gemini 3, the latest iteration of its AI platform that seeks to topple OpenAI’s dominance.
