Google’s AI Mode has emerged as the top-performing AI tool in a recent test conducted by volunteer librarians. The AI tool outshone its competitors, including Bing Copilot, ChatGPT 4 and 5, Claude, Grok, Meta AI, Perplexity, and Google itself, in terms of accuracy and reliability.
The test involved over 900 questions across various categories, including trivia, recent events, and specialized sources. The librarians used a combination of online search results, news reports, and academic databases to evaluate the performance of each AI tool.
Google’s AI Mode excelled in areas such as searching the web before providing an answer, citing sources accurately, and recognizing when information needs updating. It also performed well on questions that required specialized knowledge and nuanced critical thinking.
On the other hand, some AI tools struggled with basic tasks, such as answering simple trivia questions or identifying colors in images. Meta AI, Grok, and Perplexity were among the worst performers in these areas.
The test highlights the limitations of current AI technology and the importance of human evaluation. While AI can be a powerful tool for complex questions that require specialized knowledge, it is not yet suitable for everyday tasks that require basic critical thinking and information literacy.
Librarians emphasized the need for users to approach AI answers with a critical eye, verifying sources and checking dates before accepting information as accurate. By doing so, they can avoid relying on AI tools and instead tap into the wealth of human knowledge available through online search results and academic databases.
Source: https://www.washingtonpost.com/technology/2025/08/27/ai-search-best-answers-facts