AI Overviews was added by Google to its Search feature to take the experience to the next level. Unfortunately, what the advertising giant envisioned took a turn for the worse because numerous findings revealed that the results were telling users to ‘add some glue’ to the homemade pizza they took out of the oven. This example is just one of the several that are littered on social media, obviously raising the question of whether Google is losing its magic in improving its search engine. A company spokesperson has responded with a reply, but not everyone will appreciate it.
Company spokesperson states that the AI Overviews results that went off the rails were ‘isolated examples’
The new feature incorrectly educates users, and while it has not been developed to respond accurately to every query, generating near-accurate AI-generated responses is not AI Overviews' key strength at this time. For instance, in the second X post below, UC Berkeley geologists have recommended that individuals should eat at least one small rock per day. Looking at these results, millions of users would be in the right frame of mind to believe that the old way of searching for answers was the best way.
Really hard to believe the team working on AI Overviews didn't find a number of these issues during testing... The YMYL queries especially... Let's see how Google improves the situation, and how quickly 🙂 -> Directory Of Embarrassing Google AI Overviews
I think my favorite is… pic.twitter.com/qheWGY6LVG
— Glenn Gabe (@glenngabe) May 24, 2024
However, why is this happening, and is it entirely Google’s fault? Speaking with The Verge, the technology giant’s spokesperson Meghann Farnsworth has stated that the mistakes came from ‘generally very uncommon queries and aren’t representative of most people’s experiences.’ Google has reportedly taken action against violating its policies, with Farnsworth also mentioning that these are ‘isolated examples’ and that the company will continue working to improve its product.
Google AI Overviews. Disasters.
A thread 🧵 pic.twitter.com/NSoPfQXf6K
— Ben Alfrey (@benalfrey) May 24, 2024
In a statement provided to 9to5Google, Google states that it appreciates feedback from users, but ‘uncommon queries’ are the primary reason for these results. The latest AI-fueled saga probably has made CEO Sundar Pichai even more nervous than before.
“The vast majority of AI Overviews provide high quality information, with links to dig deeper on the web. 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. 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.”
Have you heard?
Google's AI missteps are turning its cash cow into a laughing stock.
Dive into the 13 most embarrassing AI fails that are wrecking Google AI overviews: pic.twitter.com/y2GvJJwUcK
— Vensy (@vensykrishna) May 24, 2024
To be fair, Google never claimed that AI Overviews would deliver the perfect experience, but that also does not mean that people searching for a way to keep the cheese from falling off of their pizza are recommended to apply glue to their food. It appears that OpenAI’s surprise attack with the launch of ChatGPT forced Google to deliver a rushed service, producing incredibly skewed results. So far, the Mountain View firm’s spokesperson was not very convincing with her choice of words, likely because she did not expect AI Overviews to go haywire.
