Jimmy Wales’ Warning: ChatGPT is ‘Pretty Bad’!

During his session at the opening night of the Web Summit 2023 in Portugal, Jimmy Wales, the co-founder of Wikipedia, emphasized the limitations of generative artificial intelligence models like ChatGPT. Addressing a packed audience, Wales expressed his fascination with ChatGPT as a recreational tool but also acknowledged that when used seriously, it becomes apparent that its performance is subpar. This viewpoint aligns with that of numerous researchers and entrepreneurs who remain skeptical about the capabilities of large language models (LLMs) and consider them an unreliable aid due to their errors.

During the event, which can be watched on Web Summit’s official YouTube channel, Axios senior editor Ryan Heath asked Wales, “Are generative AI companies getting the scrutiny they deserve?” Says Wales to this: “One of the things we’ve seen is that when a new technology comes out, there’s a lazy alarm.”

Highlighting the example of eBay’s early days, Wales said how people raised the alarm whenever a controversial ad appeared, but slowly people realized that it could only be reported and removed, and no illegal business was being conducted. Similarly, he also provided an example when American model and TV personality Anna Nicole Smith died in 2007 and someone briefly vandalized her Wikipedia page, leading to several news reports in which Wales questioned the lack of a Wikipedia model. Addressing its original question, Wales says that looking at generative AI is part of raising the alarm about this emerging technology.

Jimmy Wales thinks ChatGPT just isn’t good enough

In addition, Wales talked about the AI chatbot and wasn’t shy about sharing his opinions: “Looking at the birth of ChatGPT, it’s clearly not good enough. It’s great to play with, but when you actually start using it, what looked fantastic at first looks pretty bad. And it gets a lot of scrutiny, I don’t think there’s anyone in this room who hasn’t heard at least one story about ChatGPT (involving its bug).

Wales also highlighted that chatbots could be made more accountable and factually accurate, which could then be deployed across industries if information is heavily filtered and low-quality sources such as social media are avoided.