The disclosure dilemma: How AI attribution affects reactions to public health messages

Abstract

This experiment (N=1,500) examines how disclosing AI generation of public health message creation affects audience reactions. Results suggest a trade-off: up-front disclosure of AI usage significantly reduces message credibility and learning (17% less information retained) but preserves institutional credibility, but if usage of AI is revealed after the fact there is substantial damage to source trustworthiness and perceived transparency. For professional-quality content with no obvious deficits, audiences did not suspect AI involvement when undisclosed.