The AI trust paradox: how disclosure, brand authenticity and consumer mindset shape communication effectiveness

Purpose This study investigates the trust paradox in artificial intelligence (AI)-generated content, where efficiency conflicts with perceived inauthenticity. Integrating source credibility theory and congruity theory, it examines how the interplay between ad genesis (AI vs. genuine), strategic context (disclosure, brand authenticity) and consumer mindset (construal level) determines communication effectiveness. Design/methodology/approach A three-study experimental design was employed to isolate specific contingencies. Study 1 manipulated ad genesis and AI disclosure to test engagement. Study 2 examined the moderating role of brand authenticity on purchase intentions. Study 3 tested the congruity between AI message focus (hedonic vs. utilitarian) and consumer construal levels. Findings Genuine imagery generally outperforms AI-generated imagery; however, effectiveness is contingent on strategic fit. Disclosure boosts engagement for genuine ads but reduces it for AI content. AI ads enhance purchase intentions for low-authenticity brands but damage high-authenticity ones. Furthermore, hedonic AI appeals align with high-construal mindsets, whereas utilitarian appeals suit low-construal consumers. Research limitations/implications The experiments used static images in controlled online settings. Future research should test dynamic and multimodal AI content and validate the effects with real-world behavioral metrics, such as click-through and conversion rates. Practical implications AI use should fit the context. Heritage brands should avoid AI imagery, whereas challenger brands can benefit from it. Disclosure helps genuine-image ads but hurts AI-generated ads. Hedonic AI appeals suit abstract mindsets, whereas utilitarian appeals suit concrete mindsets. Originality/value This research advances advertising theory by applying a split-credibility lens to AI-generated advertising, demonstrating that effectiveness depends on source-context congruence.