From Search Engines to Answer Engines: Testing the Effects of Traditional and AI Search Engines on Belief in Misinformation

<p>Counter to the advice in many digital literacy guides, recent research has demonstrated that using a search engine to learn more about false or misleading articles can actually incorrectly increase belief in the veracity of that news (Aslett et al. 2024). As the future of search turns to AI-powered search and chatbots, will AI be able to ameliorate this problem? To investigate this question, we recruited over 4,000 participants to assess the veracity of recent news articles, and evaluate how using either a traditional search engine (TSE) or an AI search engine (AISE) affects their assessments. We find that using the AISE (Perplexity) increases belief in misinformation at almost the same rate as the TSE (Google). We also find that search strategies which perform poorly on TSEs (e.g., copying and pasting links) may actually improve the utility of AISEs. We find no difference in the rate of unreliable links served.</p>