The uncanny valley effect and the creepiness of ai generated advertising: a study on customer trust and loyalty

Author

Dr. Pradeep Kumar Patnaik, Ms. Sadhvika Reddy Mallidi


Abstract

 The aim of this s The aim of this study is to identify the relationship between advertising realism, AI disclosure, the human likeness of AI avatars, perceived creepiness of AI-generated ads and customer trust and brand loyalty. Primary data were gathered using a structured questionnaire on 103 social media users of India, and analysed by using chi-square test, one-way ANOVA and linear and multiple regression, the Baron and Kenny procedure, Sobel test and a bootstrap to check the mediation. The factors of human likeness (β = 0.76), realism (β = 0.16) and disclosure (β = 0.11) all significantly contributed to the variance in perceived creepiness, accounting for 82.6% of the variance. Contrary to the hypothesis, creepiness did not negatively influence the customer trust (β = −0.05, p = 0.614), but positively (β = 0.39, p < 0.001) influenced brand loyalty. Trust strongly predicted loyalty (β = 0.58, R² = 0.34). Contrary to the usual role of mediation, creepiness was a suppressor rather than a classical mediator: realism and human likeness increased creepiness, which, in turn, was associated with higher trust, to some extent counteracting their direct negative impact on trust. The results indicate that, for this sample, discomfort with AI-driven adverts that resemble humans has not led to any decrease in trust or loyalty, and that trust is the primary pathway to loyalty. Advertising agencies must navigate the moral and ethical landscape of AI-generated content with great caution, as consumers are growing increasingly aware of these new methods as AI tools raise the bar for the ethics and morality of advertising

tudy is to identify the relationship between advertising realism, AI disclosure, the human likeness of AI avatars, perceived creepiness of AI-generated ads and customer trust and brand loyalty. Primary data were gathered using a structured questionnaire on 103 social media users of India, and analysed by using chi-square test, one-way ANOVA and linear and multiple regression, the Baron and Kenny procedure, Sobel test and a bootstrap to check the mediation. The factors of human likeness (β = 0.76), realism (β = 0.16) and disclosure (β = 0.11) all significantly contributed to the variance in perceived creepiness, accounting for 82.6% of the variance. Contrary to the hypothesis, creepiness did not negatively influence the customer trust (β = −0.05, p = 0.614), but positively (β = 0.39, p < 0.001) influenced brand loyalty. Trust strongly predicted loyalty (β = 0.58, R² = 0.34). Contrary to the usual role of mediation, creepiness was a suppressor rather than a classical mediator: realism and human likeness increased creepiness, which, in turn, was associated with higher trust, to some extent counteracting their direct negative impact on trust. The results indicate that, for this sample, discomfort with AI-driven adverts that resemble humans has not led to any decrease in trust or loyalty, and that trust is the primary pathway to loyalty. Advertising agencies must navigate the moral and ethical landscape of AI-generated content with great caution, as consumers are growing increasingly aware of these new methods as AI tools raise the bar for the ethics and morality of advertising



Keywords

AI disclosure", "customer trust", "brand loyalty", “perceived creepiness”, “ai-advertising”



Full Text:

Download Paper PDF


References


 

Share your valuable work from Social Media Buttons