The Prospect of Embodied Intelligence in Dentistry.

Conventional artificial intelligence in dentistry is limited by a 'diagnostic-executive disconnect', functioning primarily as static 'bystander intelligence'. This review aims to delineate the technical architecture of Embodied Artificial Intelligence (EAI) in dentistry and analyze its potential to bridge this gap through a clinician-supervised closed-loop framework. We synthesized current literature from medical and engineering databases (PubMed, IEEE Xplore, Scopus) regarding the integration of multimodal perception, memory and knowledge integration, embodied reasoning and planning, and precision execution within dental systems. The review focused on literature demonstrating the transition from static diagnostic AI to active therapeutic intervention. The proposed dental EAI architecture integrates four core modules: multimodal perception, memory and knowledge integration, embodied reasoning and planning, and precision execution. Analysis reveals its clinical potential in robot-assisted implant navigation, precision endodontic microsurgery, and dynamic orthodontic monitoring. Digital twin technology is identified as a key 'decision sandbox' for prospective outcome prediction. Furthermore, the review evaluates systemic barriers such as data scarcity and algorithmic opacity, discussing Dental Foundation Models and Explainable AI (XAI) as potential approaches. However, much of the current evidence remains at the engineering, preclinical, or early clinical stage. EAI provides an emerging framework for shifting from 'bystander intelligence' to active clinician-agent partnership. By establishing a closed-loop framework of perception, planning, and execution, EAI addresses the limitations of current digital workflows. Current systems should be regarded as task-specific, clinician-supervised co-pilots rather than autonomous replacements for dental professionals, and further clinical validation is required before broader implementation. By bridging the diagnostic-executive disconnect, Embodied AI may provide a clinician-supervised framework to support more integrated dental workflows across diverse dental specialties.