SAN FRANCISCO, California – Users of the AI chatbot ChatGPT were taken by surprise when the program started generating nonsensical answers on Tuesday night, as reported by OpenAI’s status page. Screenshots of the bizarre conversations flooded social media, with the issue persisting into Wednesday morning.
OpenAI acknowledged the problem, stating that they are “investigating reports of unexpected responses from ChatGPT” on their status page. Shortly after, they updated the page, confirming that they are “continuing to monitor the situation.”
The unusual behavior of ChatGPT sparked a flurry of user reactions on Reddit and other platforms, with one user describing the AI’s responses as starting normally before “devolving into nonsense.” The responses were a mix of languages and gibberish, with some users even sharing videos and photos documenting the erratic behavior.
The cause of the bug remains unknown, and it is distinct from typical outages and heavy traffic issues. OpenAI’s status page, usually reserved for reporting such problems, now displays a warning for “unexpected responses.”
While some users have reported no issues with ChatGPT, others continue to experience the problem, suggesting that it is widespread. As of Wednesday morning, the issue was still active on OpenAI’s status page.
The company has yet to comment on the situation, leaving users and observers puzzled by the sudden malfunction of the AI chatbot. The erratic behavior of ChatGPT has left many users wondering about the underlying cause and the potential impact on the functionality of the program.