From Doubt to Action: Examining the Potential of AI Comments in Promoting Health Information Seeking on Social Media

Using a 2 (social endorsement: high vs. low) × 2 (AI comment type: approving vs. disapproving) experiment, we examine how AI-generated comments and social endorsement cues jointly influence information-seeking intentions in response to social media health misinformation. Results reveal a moderated-moderated mediation: When AI trustworthiness is high and social endorsement is low, debunking comments promote cognitive elaboration, which in turn increases information-seeking intentions. This pathway suggests that deeper cognitive processing of AI corrections makes individuals more receptive to seeking additional verification. When AI trustworthiness is low, debunking is largely ineffective. Notably, under low trustworthiness and endorsement, misinformation-supporting comments can unexpectedly elicit more elaboration, potentially triggering skepticism-driven information seeking. These findings underscore the importance of AI trustworthiness and reveal nuanced mechanisms in how individuals respond to AI-based health misinformation interventions, highlighting that cognitive elaboration serves as a critical bridge between AI interventions and behavioral intentions.