Once, when we searched for scientific information online, we learned to ask who wrote it, where it was published, and whether other sources said the same thing. But what happens when the information comes from a chatbot that speaks with confidence, summarizes sources for us, and serves up an answer that looks ready to use? In this study, we followed 30 participants from diverse educational backgrounds who were asked to address everyday dilemmas related to science, health, and food. They searched for information both on Google and with generative AI, and we examined how they decided whether to trust the information they received.
Contribution
The findings show that people change the way they check information when it comes from artificial intelligence. Instead of opening several sources and comparing them themselves, many checked whether the bot had accurately represented the information in the sources it chose to show. Instead of asking who was behind the information, they often settled for a familiar website name or for the feeling that the machine probably knew what it was doing. It was especially interesting to discover that although everyone held a conversation with the bot, almost no one used that conversation to challenge it, ask for evidence, or find out why it had chosen those particular sources. This research matters to all of us: in an age when artificial intelligence can sound like an expert, information literacy is no longer just about knowing how to search. We also need to know how to ask the bot good questions, recognize when an answer conceals uncertainty, and decide when we can trust it and when we should stop and check again.
Link to the full article: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0345300
Authors: Shakked Dabran-Zivan, Inbal Klein-Avraham, Ayelet Baram-Tsabari
Research group’s website: https://ayeletlab.net.technion.ac.il/
