Overnight on February 21st, many users reported experiencing bizarre and nonsensical responses from ChatGPT, the AI assistant created by OpenAI.
Screenshots showed the AI providing unintelligible, excessively verbose, or repetitive responses to fairly straightforward questions about topics like coding, music recommendations, and definitions of computers.
One response about computers rambled on about complex topics in a strange, philosophical manner that did not answer the question.
Users on Reddit and other forums shared their strange interactions with ChatGPT and expressed concern about its sudden erratic behavior.
It appears the issue was primarily affecting GPT-4 users who pay a monthly subscription fee for the most advanced version.
OpenAI acknowledged an unspecified problem on its status page but the exact cause remains unclear. They are continuing to monitor the situation.
Source: futurism
