AI Transparency And Consumer Judgments: A Meta-Analysis Of Disclosure And Explainability

Abstract

This study synthesises the fragmented literature on transparency in artificial intelligence (AI) to examine its relationship with consumer judgments, trust, and behavioural intentions. Drawing on a three-level mixed-effects meta-analysis of 304 effect sizes from 82 empirical articles, the study estimates the overall association between AI transparency and consumer outcomes while examining transparency predictors, outcome domains, and contextual boundary conditions. The findings indicate that AI transparency is positively associated with consumer judgments and behavioural outcomes. Notably, simple disclosure labels are statistically as effective as more complex explainability mechanisms, suggesting that greater explanatory complexity does not necessarily yield stronger consumer responses. Transparency also exhibits stronger associations with perceived system characteristics and trust than with behavioural intentions, while these associations become weaker under more demanding contextual and interface conditions. These findings extend the transparency literature by suggesting that AI transparency may function not only as a mechanism for communicating technical information but also as a heuristic signal of organisational openness. From a practical perspective, simple transparency disclosures can provide an effective foundation for building consumer trust, whereas higher-stakes AI applications may benefit from combining disclosure with human oversight and contextual explanations to encourage favourable consumer responses.