BlockBeats News, May 5th, according to The New York Times, AI startup Oumi's analysis of Google's "AI Summary" feature shows that the feature's accuracy rate is approximately 90%. Since Google processes over 50 trillion searches per year, this means it provides tens of millions of wrong answers every hour (or hundreds of thousands of inaccuracies per minute).
Oumi used the SimpleQA benchmark to analyze Google's "AI Summary" search results and found that when using Gemini 2, 85% of the results were accurate; with Gemini 3, this proportion rose to 91%.
During the analysis of the 5380 sources referenced by Google's AI Summary, Oumi found that Facebook and Reddit were the second and fourth most cited sources, respectively. When Google's AI Summary is accurate, 5% of the citations reference Facebook; when inaccurate, 7% reference Facebook.
