Understanding Older Adults’ Evaluation of AI-Generated Health Information

When older adults cannot distinguish accurate health information from misinformation within AI-generated content, they may misjudge its credibility and follow incorrect advice. To investigate how information characteristics shape older adults’ assessment of AI-generated health content, this study employed a 2 × 2 × 2 within-subject design manipulating emotional framing (fear vs. neutral), authority cues (high vs. low), and information veracity (true vs. false), with individual differences examined as exploratory variables. Seven older adults rated the credibility and adoption intention of AI-generated information and chose whether to verify it. Those verifying viewed corrections and re-rated initial information. Results showed that emotional framing and authority cues impaired participants’ ability to identify misinformation. Individuals with higher digital literacy and cognitive ability relied less on emotional framing. Findings suggest that such information characteristics make misinformation harder to identify, whereas digital literacy and cognitive ability may buffer this effect by reducing reliance on them.