At the Web Summit In Lisbon, Wikipedia founder Jimmy Wales downplayed the capabilities of viral AI chatbot ChatGPT, terming it “pretty bad” due to frequent errors.
Despite AI dominating discussions at the tech event, Wales exhibited less enthusiasm about large language models like ChatGPT. He stated such AI is still three decades away from rivaling human abilities.
This comes amid the meteoric rise and growing debate around ChatGPT, launched by OpenAI last year. Tech giants like Microsoft and Google are betting big on conversational AI.
However, Wales feels Microsoft’s partnership with OpenAI could become problematic for Google. He told the audience that for many serious purposes, ChatGPT falls short and makes a lot of mistakes.
Wales took a dig at Twitter, now rebranded as X, saying it is not a credible source for factual information. This references Elon Musk’s rocky takeover of the platform.
On Wikipedia’s potential AI integration, Wales clarified the technology won’t be used to write articles outright. Instead, it could help enhance existing articles by identifying gaps and suggesting new sources.
He noted AI tends to be more accurate on popular topics already well-documented by humans. But Wikipedia needs help on more obscure subjects where AI struggles.
“The problem with AI is that the stuff it is good at and gets correct includes content already widely written about on Wikipedia and by humans,” Wales explained.
A recent study showed ChatGPT follows clinical guidelines for depression better than doctors, sparking job loss fears. However, Wales’ comments indicate generative AI still faces limitations in matching human skills.
While AI collaboration could boost Wikipedia’s productivity, Wales remains cognizant of the technology’s current shortcomings.
