Emotional AI in Advertising: Recognizing and Responding to Consumer Affective States

Abstract

This study examines how emotional artificial intelligence can reshape contemporary advertising by recognising and responding to consumer affective states. Drawing on a systematic literature review of work in affective computing, emotion recognition, advertising, and ethical analysis, the paper synthesises evidence on the technical capabilities, applications, and risks of emotionally responsive systems. The review shows that automated emotion measures based on facial expressions, voice, physiology, and multimodal fusion can capture meaningful reactions to advertising and enhance diagnostics and pre-testing. However, these systems produce probabilistic indicators rather than direct access to “true” feelings, and their validity depends on context, data quality, and theoretical interpretation. Empirical studies reveal that emotional information is mostly used for evaluation rather than in closed loop, real time adaptation of messages. Across the literature, concerns about privacy, manipulation, and bias highlight the need for strong governance and transparency. The study outlines research and design implications for developing emotionally intelligent advertising that supports both effectiveness and consumer welfare in advertising practice.