The authenticity penalty: Why AI-enhanced performance diminishes relational value in the creator economy

Luo, Jianhong, Li, Chenyu
Electronic Markets · Springer Nature · 2026-08-24 · DOI: 10.1007/s12525-026-00930-y

Abstract

Generative AI (GenAI) is reshaping the creator economy by automating content production, yet the relational value of user-generated content relies heavily on perceived human authenticity. Does AI-enhanced performance come at a relational cost? To answer this, we leverage a quasi-natural experiment surrounding the launch of an in-app GenAI writing assistant on a major travel platform, using a comparable platform as control. Using a difference-in-differences design with intention-to-treat (ITT) estimates, we document a “double-edged sword” effect. While GenAI availability significantly boosts textual persuasiveness, it is also associated with lower audience resonance, consistent with an “authenticity penalty.” Mediation analysis provides proxy evidence that this drop is linked to the very optimization of the text, which is consistent with the activation of persuasion knowledge. Furthermore, using CLIP-based measures of text–image mismatch, we find that the penalty appears directionally stronger when polished text is less congruent with the accompanying visuals; although the strictest interaction tests are power-limited, the marginal-effects pattern is consistent with multimodal dissonance. These findings highlight the tension between automation and expressive agency, urging platforms to move beyond simple efficiency metrics toward designs that support human–AI symbiosis.

Read at publisher →

Similar articles