When strange and misleading responses to search queries created by Google’s new AI Overview feature went viral on social media last week, the company initially downplayed any issues. However, on Thursday, Liz Reid, the head of search at Google, acknowledged that there were areas that needed improvement, stating, “We wanted to explain what happened and the steps we’ve taken.”
Reid specifically mentioned two of the most viral and inaccurate AI Overview results. One result suggested eating rocks for health benefits, while another recommended using non-toxic glue to thicken pizza sauce.
Reid explained that the misinformation regarding rock eating stemmed from the AI tool misinterpreting a satirical article from The Onion as factual information. Regarding the glue on pizza suggestion, Reid attributed it to the AI tool pulling content from discussion forums that sometimes provide unhelpful advice.
Reid advised users to carefully review any AI-generated content, especially when it comes to dinner menus.
The company defended the quality of the new search feature, stating that extensive testing was conducted before the launch and that user data shows people value AI Overviews.
Reid acknowledged that the attention-grabbing mistakes were a result of widespread internet searches, some of which were deliberately misleading.
Google also addressed fake screenshots circulating on social media, clarifying that some controversial AI Overview results were not authentic. The company made technical improvements to the AI Overview feature, including better detection of nonsensical queries and reducing reliance on user-generated content.
Google will continue to monitor user feedback and make adjustments to the AI Overview feature as necessary.
