“Those AI Overviews never appeared.”
Fake News
Ever since Google rolled out its AI Overviews search feature to the masses earlier this month, it’s earned its share of bad headlines, from erroneously telling people to jump off the Golden Gate Bridge to advising them to eat rocks.
The situation had gotten so out of control, as Quartz reports, that mischief makers have been faking screenshots to make it seem like the new tech feature is doling out even more wacky responses to queries.
For example, an X user posted a purported screenshot that read, “Doctors recommend smoking 2-3 cigarettes per day during pregnancy.” This post attracted 9.9 million views.
But as other users were quick to point out, the screenshot was likely faked.
Now, Google is fighting back, accusing users of exaggerating just how braindead its new AI feature really is. Liz Reid, vice president and head at Google Search, noted the existence of a “very large number of faked screenshots” in a blog post, possibly implying that concerns over the new feature were overblown.
Nevertheless, she cautioned users to fact-check the AI tool’s dubious output for themselves, partially avoiding taking responsibility for its potentially dangerous suggestions.
“Some of these faked results have been obvious and silly,” she wrote. “Others have implied that we returned dangerous results for topics like leaving dogs in cars, smoking while pregnant, and depression.”
“Those AI Overviews never appeared. So we’d encourage anyone encountering these screenshots to do a search themselves to check,” she added.
Quick Fix
And for any actual problems or nonsensical answers that a user may encounter, Reid claimed that Google engineers are hard at work on a fix.
Because of the very odd — and not at all invented — answers that AI Overviews has generated since its launch earlier this month, the company announced it’s putting limits on where these outputs appear.
But this arguably doesn’t quite fix the issue: many users are still getting incorrect answers in AI Overviews. And whether the problem will ever truly go away remains to be seen, especially given that the company’s CEO Sundar Pichai has already admitted it doesn’t actually have a solution to what appears to be a problem inherent in the tech.
“In defense of Google’s AI Overview at its worst, it gives you erroneous laughable and even dangerous misinformation,” tweeted New York Times editor Adam Sternbergh. “But at its best, it gives you the first few lines of a Wikipedia entry.”
More on Google Search: Google Admits Its AI Search Feature is a Dumpster Fire, Says It Will Scale Back the Tool
Share This Article