ChatGPT users were amazed by the AI bot that answered their queries in Welsh.
The innovative technology is impressive, but it is still not without its flaws, as users discovered when it began translating their English questions and answers into Cymraeg.
The Financial Times was the first to report on the problem, which users have been complaining about for some time.
This only appears to happen when using the voice feature, where users say their requests out loud rather than typing them in. The company told the FT it was aware of an issue with its voice transcription system, Whisper.
Last month, the Welsh Government entered into a new “data partnership” with OpenAI, the parent company of ChatGPT, to help it understand the language and improve the usefulness of AI for Welsh speakers.
The Welsh Government recently agreed a new ‘data partnership’ with OpenAI
![ChatGPT users are left lost in translation when the AI program starts responding to them in Welsh 7 A recent study found that people find jokes produced by ChatGPT funnier than humans.](https://whatsnew2day.com/wp-content/uploads/cwv-webp-images/2024/08/1723805125_30_ChatGPT-users-are-left-lost-in-translation-when-the-AI.jpg.webp)
A recent study found that people find jokes produced by ChatGPT funnier than humans.
AI technology continues to develop at a rapid pace.
Just a few weeks ago, a recent study found that people find jokes produced on ChatGPT funnier than those written by humans.
Participants in the University of Southern California study did not know who wrote the jokes before they said which ones they preferred.
They overwhelmingly voted in favor of AI jokes: 75 percent said they were funnier than humans.
Meanwhile, a leading US medical body is encouraging doctors to use it to free up their time.
A study by the American Academy of Family Physicians looked at how well AI models can interpret and summarize complicated medical studies, which doctors are encouraged to read to stay up to date on the latest research and treatment advances in their field.
They found that the chatbot was accurate 98 percent of the time, providing doctors with quick, accurate summaries of studies across a range of specialties, from cardiology and neurology to psychiatry and public health.
(tags to translate)dailymail