ChatGPT has quickly become one of the most popular AI tools in the world, but a recent bug has caused the popular chatbot to generate unexpected responses.
ChatGPTis one of creatorOpenAI‘s most popular AI applications. But, this is a far cry from theirnew “Sora” AI video generationsoftware. After undergoing refinements and even a change to the advancedGPT-4 system, the chatbot can still generate some questionable responses, without the need forjailbreaksof any kind.
The bug was first spotted on the ChatGPTsubreddit, which is full of posts reporting “nonsense” responses. One post is titled “This felt like watching someone slowly go insane”, where the response starts off strong, but slowly begins to falter into absolute nonsense. It appears that these hallucinations cause the language-learning AI to just start using phrases without context.
AI “hallucinations” are pretty commonplace amongst language learning modules, and even though ChatGPT might be one of the most advanced of them out there, these responses prove that there’s still more work to be done.
Aviral storyfrom tech columnist Ed Zitron further explains that some companies might be reaching the “upper limits” of what is possible using AI, and how accurate outputs can really be.
OpenAI is aware of the widespread issues
OpenAI has acknowledged the current errors with ChatGPT on itsservice status page. The incident note reads: “We are investigating reports of unexpected responses from ChatGPT.” and that the company has identified the issue and is working on a fix.
As of the time of writing, the issue has not yet been resolved. However, OpenAI has a pretty good track record of fixing errors and loopholes to create unexpected responses, such as the now-infamous “Grandma” exploit.
Sayem Ahmed was Dexerto’s Tech Editor in the UK team, leading hardware coverage globally. Sayem is an expert in all things Nvidia, AMD, Intel, and PC components. He has over 10 years of experience, with bylines at Eurogamer, IGN, Trusted Reviews, Kotaku, and many more.
