How privacy communication formats shape sensitive data disclosure in AI applications

Purpose This study aims to examine how the privacy communication format (text, infographic or audio) affects users’ willingness to provide sensitive information and their intention to use health and sports apps, including artificial intelligence (AI)-enabled sports applications, and whether subjective privacy literacy and privacy concerns moderate these effects. This paper highlights both the positive and negative aspects of AI-driven service design. Design/methodology/approach In a between-subjects online experiment (N = 396), participants were randomly assigned to 1 of 3 formats that presented the same privacy content. Willingness to provide personal information, intention to use, subjective privacy literacy and privacy concerns were measured using validated scales. Moderated mediation analysis was conducted using Hayes' PROCESS macro (Model 9). Findings Audio messages reduced willingness to provide personal information compared to text, while infographics did not differ significantly from text. Willingness to provide personal information mediated the relationship between format and intention to use in the audio-versus-text comparison. Subjective privacy literacy attenuated the negative impact of audio, whereas privacy concerns did not. The results show that, in this setting, text-based privacy communication is associated with greater disclosure readiness than audio, while no significant differences emerged between infographic and text formats. Practical implications Organizations may benefit from treating privacy communication format as more than a compliance formality, especially in sensitive AI-enabled service contexts. Allowing users to choose their preferred format or tailoring messages to their literacy levels can mitigate the drawbacks of AI-enabled disclosure processes and support the ethical and effective deployment of technology. Originality/value Based on the communication privacy management and media richness theories, this study contributes to understanding how the presentation format of privacy notices influences disclosure boundaries and adoption in AI-enabled services. It also shows how individual differences in privacy literacy shape users' responses to different communication formats, particularly between audio and text.