Trust in Generative AI for Health Information Consumption and the Effect of Learned Dependency: Randomized Controlled Experimental Study
Two randomized controlled experiments (338 college students, then a replication with 563 Amazon Mechanical Turk participants) using a 2 x 2 between-participants design that manipulated whether AI-generated health information was correct or incorrect and whether critical text was highlighted. Participants evaluated the AI output alongside the source text; trust was measured with a multi-item scale and habitual reliance on generative AI with a validated learned-dependency measure.
Publisher
Journal of Medical Internet Research (JMIR Publications); University of Arizona
Published
16 Sept 2026
Added
today
Key Findings
- Participants trusted correct AI-generated health information more than incorrect information in both experiments
- Learned dependency on generative AI was positively associated with trust overall
- The accuracy by dependency interaction was negative and significant in both experiments: participants with higher habitual dependency were less sensitive to inaccuracy
- Highlighting critical information in the text had no effect on trust and did not moderate the dependency effect
Methodology Notes
Two 2 x 2 between-participants experiments; n=338 students and n=563 MTurk workers; linear regression with interaction terms; self-reported learned dependency; single-exposure evaluation task rather than conversational use. JMIR 2026;28:e98326, published 16 September 2026; PubMed indexes it as a randomized controlled trial. The JMIR site does not serve this host directly; verified from the PubMed record (42747973) and the article rendered through a reader proxy.
Sources
Topics
Authors
Ahmed, Arif, Leroy, Gondy, Sachdeva, Agrim, Harber, Philip, Rains, Stephen, Youn, Seokjun, Barai, Prosanta
Tags
Cite This
APA
Ahmed, Arif et al. (2026). Trust in Generative AI for Health Information Consumption and the Effect of Learned Dependency: Randomized Controlled Experimental Study. Journal of Medical Internet Research (JMIR Publications); University of Arizona. https://www.jmir.org/2026/1/e98326