An Integrative Cognitive Model for Multisensory Design: Benefits and Risks of AI-Personalization

Roxane L. Bartoletti^1,2*Xavier Corveleyn^2*

• ¹Universite de Nimes, Nimes, France
• ²Universite Cote d'Azur Laboratoire d'Anthropologie et de Psychologie Cliniques Cognitives et Sociales, Nice, France

In this paper, we examine the interplay between multisensory environments, cognitive performance, and artificial intelligence (AI)-enabled personalization. We propose an integrative cognitive model to better understand how the personalization of the sensory environment influences behavior, emotion, and cognition, drawing upon the Cognitive Capacity Hypothesis, Load Theory, Distraction-Conflict Theory, and the Strength and Vulnerability Integration model. Our integrative model delineates how the characteristics of the individual, the task, and the sensory stimuli interact through arousal modulation. Based on recent conceptual and empirical studies, this model proposes that (1) optimal arousal could improve distractor inhibition and task-focusing, (2) metacognitive misjudgments could lead individuals to select suboptimal sensory environments, and (3) aging alters sensory processing efficiency, necessitating tailored approaches. Within this theoretical proposition, we argue that sensory stimuli modulate arousal and available cognitive capacity, thereby influencing cognitive performance. Thus, when expanding to AI, personalized uni-and multisensory environments could demonstrate both benefits (e.g., enhanced attentional states, therapeutic applications) and risks (e.g., privacy erosion, metacognitive biases). Empirical evidence suggests that preferred background music can reduce mind-wandering, while olfactory stimuli, though underutilized in Western societies, hold untapped potential due to their strong links to memory and emotion. Whereas AI-personalized sensory environments open new perspectives into user experiences and therapeutic approaches (e.g., VR, music therapy, multisensory environment), they raise ethical concerns as the use of algorithms may polarize preferences and exploit behavioral data. Future research should address ethical AI design while leveraging cross-modal correspondences to enhance cognitive, emotional, and behavioral outcomes. Overall, this integrative model proposes an integrative framework by gathering all essential elements for creating a meaningful and coherent multisensory environment, which could be applied to researchers, artists, or marketers.