The media literacy dilemma: can ChatGPT facilitate the discernment of online health misinformation?

Online health misinformation carries serious social and public health implications. A growing prevalence of sophisticated online health misinformation employs advanced persuasive tactics, making misinformation discernment progressively more challenging. Enhancing media literacy is a key approach to improving the ability to discern misinformation. The objective of the current study was to examine the feasibility of using generative AI to dissect persuasive tactics as a media literacy scaffolding tool to facilitate online health misinformation discernment. In a mixed 3 (media literacy tool: control vs. National Library of Medicine [NLM] checklist vs. ChatGPT tool) × 2 (information type: true information vs. misinformation) × 2 (information evaluation difficulty: hard vs. easy) online experiment, we found that using dissecting persuasive strategies of ChatGPT can be equally effective when compared with the NLM checklist, and that information type was a significant moderator such that the ChatGPT tool was more effective in helping people identify true information than misinformation. However, the ChatGPT tool performed worse than control in terms of helping people discern misinformation. No difference was found in terms of perceived usefulness and future use intention of the ChatGPT tool and the NLM checklist. The results suggest that more interactive or conversational features might enhance usefulness of ChatGPT as a media literacy tool.