A Novel Paradigm in Consumer Research Driven by Generative and Agentic AI：Multimodal Perception，Brain-inspired Simulation，and Autonomous Evolution

This article systematically reviews the paradigm shifts in consumer research in the era of digital intelligence. Focusing on the challenge of dynamically modeling the complexity of consumer behavior，it proposes an AI-driven “three-stage” new paradigm powered by large models and intelligent agents—from achieving multimodal consumer perception via discriminative AI，to performing brain-inspired cognitive and decision-making simulations through generative AI，and ultimately progressing to agentic AI with autonomous evolution capabilities，including self-learning and environmental interaction. This paradigm surpasses the conventional “classification—recognition—rule-matching” approach by constructing a dynamic “consumer world model” and enabling cross-contextual generalized modeling of consumer behavior through multi-agent collaboration to simulate market group effects and strategy intervention outcomes. Emphasizing a multidisciplinary integration from a management perspective，the study explores key scientific issues such as AI-based modeling of consumer behavior，comprehensive AI embedding，and AI for social sciences，while addressing challenges like data ethics，model biases，and validation mechanisms. This paradigm promises methodological innovations for resolving issues such as predicting behavioral heterogeneity and dynamically optimizing marketing strategies，thereby advancing consumer research from “passive insights” to “active evolution” and opening new avenues for sustainable consumption practices and marketing science innovations in human-AI collaboration.