In Short:
Google’s new AI Overview feature produced bizarre and incorrect search results, causing a social media uproar. However, Google defended the technology but admitted areas of improvement were needed. The errors were attributed to lack of sources and misinterpretation from satirical websites. Google made technical improvements to the AI Overview feature to avoid nonsensical responses and rely less on user-generated content. The company will continue to monitor feedback and make necessary adjustments.
Google Acknowledges Issues with AI Overview Feature
After bizarre and misleading responses to search queries surfaced on social media due to Google’s AI Overview feature, the company’s head of search, Liz Reid, admitted that improvements were needed. In a statement on Thursday, Reid addressed the errors and steps taken to rectify them.
Incorrect Results
Two widely shared inaccurate results included suggestions to eat rocks for health benefits and use non-toxic glue to thicken pizza sauce. Reid explained that the algorithm mistakenly treated satirical content from The Onion as factual information, leading to these misleading responses.
Error Correction
Reid attributed some errors to misinterpretation of content from online forums and emphasized the importance of verifying AI-generated information before acting on it, especially in the case of crafting dinner menus.
Quality Assessment
She urged users not to judge the entire AI Overview feature based on isolated screenshots, citing extensive testing conducted before the launch. Reid highlighted user preferences for AI Overviews and defended the technology’s overall value.
Improvements Implemented
Google made technical enhancements to the AI Overview feature, including better query detection, reduced reliance on user-generated content, and stricter controls on sensitive topics like health. Reid assured continued monitoring and adjustments based on user feedback.
