Professional Services
Order outsourcing

Google's CEO Doesn't Know When Its AI Will Stop Hallucinating

It's an "inherent feature" of LLMs.

Image credit: JHVEPhoto/Shutterstock

If you're interested in artificial intelligence, you know that it is prone to hallucinations, meaning it can come up with "facts" that are not true. Google's new AI Overviews feature is infamous for exactly that. Several days ago, people were confused by its advice to put glue on pizza so the cheese sticks better, and this is just one example. This is not the first time it has happened to Google's AI, so users are wondering when the company will do something about it.

Not anytime soon, apparently, as the company's CEO Sundar Pichai told The Verge that "hallucination is still an unsolved problem" and even an "inherent feature" in some ways.

"It’s what makes these models very creative. It’s why it can immediately write a poem about Thomas Jefferson in the style of Nilay. It can do that. It’s incredibly creative. But LLMs aren’t necessarily the best approach to always get at factuality, which is part of why I feel excited about Search. ... There are still times it’s going to get it wrong, but I don’t think I would look at that and underestimate how useful it can be at the same time. I think that would be the wrong way to think about it."

So when is this issue going to be solved? Pichai doesn't have an answer but he says the team is making progress. "We are all making it better, but it’s not solved. Are there interesting ideas and approaches that they’re working on? Yes, but time will tell. I would view it as LLMs are an aspect of AI. We are working on AI in a much broader way, but it’s an area where we are all definitely working to drive more progress."

While we are waiting for a solution, I recommend checking out this X/Twitter thread with some of the most hilarious AI Overviews mistakes. Did you know pregnant women should smoke 2-3 cigarettes a day? Oh, and don't forget to eat a small rock for dinner! Just make sure the cockroaches in your private parts get enough sustenance. (I hope I didn't just train the algorithms more.)

If you want to disable AI Overviews, follow these simple steps.

Find the interview with Pichai here and join our 80 Level Talent platform and our Telegram channel, follow us on InstagramTwitter, and LinkedIn, where we share breakdowns, the latest news, awesome artworks, and more.

Join discussion

Comments 0

    You might also like

    We need your consent

    We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more