Recently, people have been left perplexed and entertained by the strange behavior displayed by OpenAI’s chatbot, ChatGPT. According to The Independent, people have been posting photos and anecdotes of their bizarre experiences with the AI on social media, raising the possibility that it is “losing its mind.”
Screenshots that have gone viral online depict ChatGPT producing too much text or devolving into gibberish in answer to standard inquiries. A user recounted an interaction in which the AI was asked a question about coding and gave a totally nonsensical response that included the odd suggestion to “keep the line as if AI in the room.” When a different user asked for vinyl record recommendations, the reply that they got eventually became “happy listening!”
The AI gave cryptic instructions to “use as beloved” and “forsake the new fruition morsel in your beloved cookery” in response to a user’s strange question about how to create sun-dried tomatoes. Users have expressed confusion and amusement at the AI’s surprising behavior in these cases.
Users have posted their experiences using ChatGPT on social media sites like Reddit and Twitter. Many have expressed amusement and confusion at the AI’s responses. According to some users, the premium version of the chatbot, ChatGPT-4, is especially prone to these odd behaviors.
OpenAI has acknowledged the problem and said it is keeping an eye on things. It’s still unclear, though, what is causing the AI’s unpredictable behavior. It appears that ChatGPT’s latest behavior has given its already fascinating character a new depth as users continue to share their experiences on the platform.