Cognitive Offloading to Generative AI: A Theoretical Framework for Memory, Metacognition, and Consumer Decision-Making

As generative artificial intelligence (Gen-AI) increasingly mediates product search, evaluation, and choice, consumer cognition is becoming distributed across human–AI systems. This shift challenges a largely unexamined assumption in advertising theory: that persuasion operates through cognitive elaboration within the individual consumer. Drawing on cognitive offloading research from cognitive psychology, we develop the CogOffload framework to explain when consumers delegate cognitive work to Gen-AI and what consequences follow. A seven-stage delegation continuum maps increasing levels of offloading, and a value-based cost-benefit framework identifies its antecedents and the moderators that shape how consumers weigh them. We propose three clusters of outcomes. First, offloading weakens memory for brands and product attributes while strengthening memory for AI-generated summaries. Second, it inflates metacognitive self-esteem and decision confidence, producing AI-induced overconfidence that reduces further information-seeking. Third, effects on decision quality depend on AI reliability: when AI is unreliable, verifying AI recommendations imposes cognitive and emotional costs but can restore impaired memory and recalibrate inflated confidence. Seven propositions and three long-term conjectures extend these effects to consumer autonomy, competence erosion, and brand loyalty. Together, the CogOffload framework positions advertising in the era of Gen-AI as operating within a distributed system of human and artificial cognition.