Google’s new AI search tool putting out wacky results
- Google rolled out new AI search tool this month
- Users got inaccurate answers to questions because of it
- Search engine says it's trying to get rid of errors
17 April 2024, USA, Palo Alto: View of the Google headquarters during the visit of North Rhine-Westphalia’s Minister President Wüst. Photo: Stella Venohr/dpa (Photo by Stella Venohr/picture alliance via Getty Images)
ovp test
Male shot by female at Shreveport apartment
Class to create biodiverse backyard
Rules for outbursts at Caddo School Board Meeting
Testing widget old system
(NewsNation) — If you’re looking to ask Google a question, you might not want to trust the first answer you see.
This month, Google unveiled AI Overviews, which puts summaries of popular answers to questions on top of search results. However, things quickly went awry.
One user wrote that Google’s new AI tool cited a satirical article by The Onion claiming geologists suggest people eat one small rock a day in response to their query. Another said it recommended putting glue on their pizza — based on a Reddit comment that is over a decade old. One Associated Press reported asked Google if cats have ever been on the moon — to which AI Overviews answered that they have.
“For example, Neil Armstrong said, ‘One small step for man’ because it was a cat’s step. Buzz Aldrin also deployed cats on the Apollo 11 mission,” Google said.
Of course, all of this is false, which many online were quick to point out. Now, Google is “racing” to manually disable AI Overviews for specific searches, The Verge reports.
While many have had fun with the system, making memes and posting their own amusing search results, experts are sounding the alarm.
Those who talked to the AP said it could perpetuate bias and misinformation, or even put people in danger when they look for help in an emergency.
“Google’s AI system is not smart enough to figure out that this citation is not actually backing up the claim,” Melanie Mitchell, an AI researcher at the Santa Fe Institute in New Mexico, said in an email to the AP. “Given how untrustworthy it is, I think this AI Overview feature is very irresponsible and should be taken offline.”
In a statement to the outlet, Google said it’s taking “swift action” to fix errors and develop “broader improvements,” though it maintains that the “vast majority of AI Overviews provide high-quality information, with links to dig deeper on the web.”
“The vast majority of AI Overviews provide high-quality information, with links to dig deeper on the web,” Google said a written statement. “Many of the examples we’ve seen have been uncommon queries, and we’ve also seen examples that were doctored or that we couldn’t reproduce.”
The Associated Press contributed to this report.
