Emotional Responses to AI-Powered Personalised Advertising: The Role of Perceived Empathy and Social Cognition in Consumer Decision-Making

The rapid proliferation of artificial intelligence (AI) in digital advertising has fundamentally transformed how brands communicate with consumers, shifting from generic mass messaging toward highly personalised, emotionally targeted experiences. Despite growing interest in AI-driven marketing, limited empirical research has examined how consumers’ socio-emotional processing mechanisms, particularly perceived empathy and social cognition, mediate the relationship between AI-powered ad personalisation and downstream consumer decision-making outcomes. This study addresses this gap by investigating the emotional and cognitive responses triggered by AI-personalised advertising among Romanian consumers. Using a quantitative survey design, data were collected from a sample of 234 adult respondents (18–65 years) in Romania, broadly aligned with key Romanian demographic distributions across age, gender, and residential area. Structural equation modelling using the Partial Least Squares (PLS-SEM) approach was employed to test the proposed conceptual model, which integrates constructs of AI-powered ad personalisation, trust in AI, perceived AI empathy, emotional arousal, cognitive elaboration, social cognition, consumer engagement, and purchase intention. The results reveal that perceived empathy toward AI-generated advertising positively influences emotional arousal and cognitive elaboration, which in turn significantly predict consumer engagement and purchase intention. Trust in AI emerged as a critical sequential mediator, while social cognition moderated the personalisation-to-trust pathway. The study yields a validated marketing model that captures the socio-emotional dynamics underlying consumer responses to AI advertising. These findings contribute to the theoretical understanding of human–AI interaction through a social cognition and emotions lens, while offering practical implications for the design of emotionally intelligent, AI-driven advertising strategies. Limitations and future research directions are discussed.