Remember that time you asked Google something and its fancy new AI Overviews feature told you to slather glue on your pizza? No? Well, you’re not alone. But guess what? It turns out these wildly wrong answers might be a permanent fixture, according to Google CEO Sundar Pichai himself.

In a recent interview with The Verge, Pichai admitted that the strange and incorrect information spewed by AI Overviews is a built-in feature of the underlying technology, large language models (LLMs). Here’s the kicker: he also said there’s no real fix in sight. So, buckle up for more nonsensical pronouncements like the one claiming 13 US presidents went to the University of Wisconsin-Madison (spoiler alert: that’s very wrong).

But wait, there’s more! Despite the, ahem, “hallucinations” being produced by AI Overviews, Pichai downplays the seriousness of the errors. He argues that while the AI might be wrong sometimes, it’s still a valuable tool. Think of it as a quirky companion, not a reliable source.

Now, some might find that logic a bit…cheesy (see, we can make pizza jokes too!). Online critics are certainly not buying it. Many are pointing out how these errors are tarnishing Google Search’s already troubled reputation for delivering less-than-stellar results.

AI experts are also chiming in, with some like Britney Muller, an AI consultant and SEO expert, questioning Google’s priorities. Muller argues that Google is focusing on gimmicks like AI Overviews when it could be developing AI for more important applications.

So, what does this all mean? Well, Pichai’s comments suggest that Google, and by extension the entire internet, might be in for a bumpy ride with AI. AI Overviews might be here to stay, flaws and all. The question remains: will users learn to trust a system that tells them to glue their cheese to the pizza, or will they ditch it for a more reliable source of information?