Testing suggests Google's AI Overviews tells millions of lies per hour

Google's AI Overviews generate millions of falsehoods per hour, a damning indictment of AI search reliability. Independent audits show 1 in 20 Overviews contain factual mistakes, a rate three times higher than the average AI chatbot and significantly worse than Bing's GPT-4-powered summaries, which error in only 1 in 60 results. This outperforms #2 in sheer misinformation volume, with billions of daily queries translating to hundreds of thousands of errors per minute. The data, based on systematic testing of 10,000 prompts, underscores a critical flaw: the system hallucinates at a 5% clip, compared to just 1.7% for leading alternatives. For users seeking accurate information, this margin of error is unacceptable.