The problem arises because artificial intelligence has no clue what any text symbols mean. You can train it to generate words based on a given input, but it will not comprehend the meaning behind those words.
"Generative AI systems like Google's AI Overview or ChatGPT don't know what's true. They just repeat what they find frequently on the web," Toby Walsh, professor of AI at UNSW Sydney and author of 'Faking It! Artificial Intelligence in a Human World,' told Bored Panda.
These overviews are one of the most dramatic examples of the limitations behind today's AI technologies, which Elon Musk said will outsmart humans by 2026.
Walsh explained, "Obscure topics are likely to give generative AI the most difficulty. No one had said on the web not to eat rocks because it was so obviously a bad idea. But there was one satirical article from the Onion that did suggest you should because of their 'good' mineral content. Unfortunately, AI can't tell satire from truth."
#7
#8
A Google spokesperson told the BBC these were "isolated examples" and insisted the feature was generally working well.
"The examples we've seen are generally very uncommon queries, and aren't representative of most people's experiences," the company's statement said.
"The vast majority of AI overviews provide high-quality information, with links to dig deeper on the web."
#11
Google also said, "We conducted extensive testing before launching this new experience, and as with other features we've launched in Search, we appreciate the feedback."
"We're taking swift action where appropriate under our content policies, and using these examples to develop broader improvements to our systems, some of which have already started to roll out."
#14
#15
But comedy isn't what we use its search engine for. "It's a risky strategy for Google to roll something like this out," Professor Walsh added.
"We are used to trusting Google to give us correct search results. Google should never forget we're only one click away from finding a new search engine."
#18
