ChatGPT user says AI bot initiated a conversation with him unprompted

  • ChatGPT appeared to initiate a conversation with a user
  • The AI chatbot asked the user a question unprompted
  • OpenAI says the issue has now been addressed 

Published on Sep 24, 2024 at 11:23 AM (UTC+4)
by Claire Reid

Last updated on Sep 24, 2024 at 3:04 PM (UTC+4)
Edited by Tom Wood

ChatGPT appeared to initiate a conversation with a user completely unprompted due to a bug. 

For those of you familiar with the AI software, you’ll know that ChatGPT responds to prompts put in by a human user. 

However, in a science-fiction-like turn of events recently, one user claims the AI bot asked him a question unprompted. 

OpenAI has since confirmed that the questioning was a bug and not a new feature of ChatGPT.

DISCOVER SBX CARS: The global premium car auction platform powered by Supercar Blondie

ChatGPT asked a user a question unprompted

AI technology has been moving at an alarming rate in recent years. 

Back in May, OpenAI launched its new GPT-4o AI model which was able to talk with other instances of itself and even sing, while Elon Musk has already predicted when he thinks AI will become superior to human intelligence. 

Meanwhile, over the weekend, an exchange between ChatGPT and a Reddit user went viral – as it appeared to show the first time the software had asked a question unprompted. 

In the exchange, which was shared on Reddit by a user with the handle SentuBill, the chatbot asked: “How was your first week at high school? Did you settle in well?”

“Did you just message me first?” SentuBill asked. 

“Yes, I did!” ChatGPT replied. “I just wanted to check in and see how things went with your first week of high school. If you’d rather initiate the conversation yourself, just let me know!”

OpenAI says it has ‘addressed’ the issue 

It wasn’t long before fellow ChatGPT users reported similar experiences, with one person saying they had asked about some health symptoms last week and ‘this week it messages me asking me how I’m feeling and how my symptoms are progressing’. 

Some users said they hoped the feature was real and looked forward to seeing it for themselves, while others were more skeptical of the exchange and suggested it could have been faked.

However, OpenAI has since addressed the speculation saying that the ‘issue’ was now resolved. 

“We addressed an issue where it appeared as though ChatGPT was starting new conversations,” OpenAI told Futurism in a statement. 

“This issue occurred when the model was trying to respond to a message that didn’t send properly and appeared blank. As a result, it either gave a generic response or drew on ChatGPT’s memory.”

So it seems like you might have to wait a little longer until an AI bot decides to ask how your day is going.

# Tags - Tech


user

Claire Reid

Claire Reid is a journalist who hails from the UK but is now living in New Zealand. She began her career after graduating with a degree in Journalism from Liverpool John Moore’s University and has more than a decade of experience, writing for both local newspapers and national news sites. Across her career she's covered a wide variety of topics, including celebrity, cryptocurrency, politics, true crime and just about everything in between.