Words that sell: Examining the impact of AI-generated language on customer trust, engagement, and business value creation

Abstract

This study employed a mixed-methods experimental design to investigate consumer responses to human-authored versus AI-generated content, with and without disclosure of AI involvement. Quantitative analyses revealed that human-authored content consistently achieved the highest trust and engagement scores, while AI-generated content with disclosure outperformed undisclosed AI content, demonstrating the importance of transparency in mitigating algorithm aversion. Regression and mediation analyses indicated that perceived authenticity significantly predicts trust and engagement, which in turn mediate purchase intention, confirming the psychological mechanisms underlying consumer decision-making. Qualitative interviews further revealed emotional and ethical responses to AI-generated content, including feelings of moral disgust toward undisclosed AI authorship and appreciation for personalized, transparent interactions. These findings collectively emphasize that ethical deployment of AI, particularly with disclosure and attention to perceived authenticity, can enhance customer engagement, trust, and loyalty. Moreover, this study will bring to focus the increasing need to have a human-in-the-loop approach, in which AI is applied to enhance and not completely displace human creativity. With the focus on transparency, brands will be able to move the potentially deceptive automation into the partnership in value creation.