Phrases such as ‘ChatGPT has gone crazy’ or ‘ChatGPT has lost it’ have been going around the internet for the whole day but is it true?
Netizens have reportedly been posting online about the artificial intelligence chatbot ChatGPT giving nonsensical and in some cases concerning responses when users would make use of the application.
Developed by OpenAI, the chatbot is based on a large language model and has become quite popular. But its recent erratic behaviour has certainly got people talking.
Did ChatGPT Lose It?
Apparently, the behaviour was first noticed on the ChatGPT Subreddit, where a user posted a video of the AI glitching.
ChatGPT 4 going crazy? The AI is trying to tell me something…
byu/Looksky_US inChatGPT
Another post “This felt like watching someone slowly go insane” showed the AI giving regular responses initially but then going into sending nonsensical gibberish.
A different post titled “Any idea what caused this?” showed the user asking the AI “synonym for overgrown” and ChatGPT going into a look of “A synonym for ‘overgrown’ is ‘overgrown'” over and over repeatedly to which the user then replied, “what the f*ck”.
Read More: Recruiters Could Reject You Immediately If Your CV Is Written By ChatGPT, Claims Study
Throughout the day, users have been posting photos and videos online showing how ChatGPT glitches, this included getting caught in a loop, constantly repeating its response, jumping from language to language, and in some cases giving absolute gibberish as a response.
A Reddit user replying to a thread posted how when they were talking about the Jackson family of musicians with the chatbot it then proceeded to send “Schwittendly, the sparkle of tourmar on the crest has as much to do with the golver of the ‘moon paths’ as it shifts from follow” among other stuff as a response.
One user while talking about “jazz albums to listen to on vinyl” as per Independent saw the chatbot get caught in a loop of sending “Happy listening!” and other gibberish.
OpenAI soon enough responded to these glitches, writing “We are investigating reports of unexpected responses from ChatGPT.”
Other wacky responses had ChatGPT suddenly starting using Spanglish, babbling, giving disturbing threats like “I’m in the room with you right now, lmao” and when one was talking about “solutions to math problems or JavaScript bug fixes” the chatbot spammed long paragraphs of text that was completely random and not related to the prompt at all.
Within just a few minutes the page was updated that they had identified the cause of the glitch and that “The issue has been identified and is being remediated now,” however, OpenAI has not publically said what the issue was or why it was happening.
As per reports, some suspect this to be due to ‘AI hallucinations’ a common phenomenon in language learning modules, where ChatGPT just started giving out phrases without any proper context to them.
Some others believe that considering ChatGPT is based on large language models that work more as predictive text, the bug could be something similar to a social media trend seen often where people are told to type a word and then form sentences out of the next work suggestion.
Image Credits: Google Images
Feature image designed by Saudamini Seth
Sources: The Indian Express, Business Insider, India Today
Find the blogger: @chirali_08
This post is tagged under: ChatGPT, ChatGPT crazy, ChatGPT openai, openai, ChatGPT app, ChatGPT news, ChatGPT ai, ai, openai news, ChatGPT malfunction, ChatGPT crazy, ChatGPT crazy responses, ChatGPT going crazy, ChatGPT goes crazy, ChatGPT gone crazy
Disclaimer: We do not hold any right, copyright over any of the images used, these have been taken from Google. In case of credits or removal, the owner may kindly mail us.
Other Recommendations:
Using ChatGPT As Your Therapist: Read This Before You Proceed
