Polite AI mitigates user susceptibility to AI hallucinations
- 28 November 2024
- journal article
- research article
- Published by Taylor & Francis in Ergonomics
- Vol. 68 (10) , 1735-1745
Abstract
With their increased capability, AI-based chatbots have become increasingly popular tools to help users answer complex queries. However, these chatbots may hallucinate, or generate incorrect but very plausible-sounding information, more frequently than previously thought. Thus, it is crucial to examine strategies to mitigate human susceptibility to hallucinated output. In a between-subjects experiment, participants completed a difficult quiz with assistance from either a polite or neutral-toned AI chatbot, which occasionally provided hallucinated (incorrect) information. Signal detection analysis revealed that participants interacting with polite-AI showed modestly higher sensitivity in detecting hallucinations and a more conservative response bias compared to those interacting with neutral-toned AI. While the observed effect sizes were modest, even small improvements in users' ability to detect AI hallucinations can have significant consequences, particularly in high-stakes domains or when aggregated across millions of AI interactions.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- To protect science, we must use LLMs as zero-shot translatorsNature Human Behaviour, 2023
- The role of politeness in human–machine interactions: a systematic literature review and future perspectivesArtificial Intelligence Review, 2023
- Someone out there? A study on the social presence of anthropomorphized chatbotsComputers in Human Behavior, 2022
- Enhancing chatbot effectiveness: The role of anthropomorphic conversational styles and time orientationJournal of Business Research, 2021
- From ‘automation’ to ‘autonomy’: the importance of trust repair in human–machine interactionErgonomics, 2018
- Evaluating Attitudes and Experience With Emerging Technology in Cadets and Civilian UndergraduatesMilitary Psychology, 2017
- Linguistic politeness and face‐work in computer mediated communication, Part 2: An application of the theoretical frameworkJournal of the American Society for Information Science and Technology, 2008
- A Note on ROC Analysis and Non-Parametric Estimate of SensitivityPsychometrika, 2005
- Development of NASA-TLX (Task Load Index): Results of Empirical and Theoretical Research1988
- Politeness1987