EAI: the clinician-supervised co-pilot during clinical procedures
A dynamic closed-loop system of perception, planning and execution
Combining artificial intelligence (AI) with sensory technologies and robotics – so-called embodied artificial intelligence (EAI) – could serve as a clinician-supervised co-pilot during clinical procedures.
Drawing on dental, medical and engineering literature, researchers examined the technologies needed to enable EAI systems and their potential clinical applications. They organised their review around four interconnected functions: gathering information from the clinical environment, integrating it with patient information and existing knowledge, planning interventions and guiding their physical execution.
An key requirement for EAI is the ability to collect reliable information in the oral cavity. The review highlighted research into force and haptic sensing technologies that enable robotic systems to detect changes in mechanical resistance during implant procedures, potentially supporting anatomical landmark detection and assessment of primary stability.
Other developments include robotic systems using optical tracking to compensate for simulated patient movements and orthodontic aligners that measure occlusal forces for treatment monitoring.
EAI systems would also need to draw on patient histories and existing clinical knowledge to inform decisions. The review highlighted research into AI models capable of processing different types of dental information, as well as datasets containing findings from repeated examinations and follow-up visits.
The researchers also discussed approaches developed for general robotic surgery that could enable EAI systems to learn from previous procedures and apply that knowledge to clinical decisions.
In AI-assisted treatment planning, studies have investigated systems that use bone anatomy and prosthetic requirements to generate implant placement plans, including for zygomatic implant procedures. The researchers also discussed how digital twins could be used to simulate different treatment scenarios and anticipate possible outcomes.
Regarding the physical execution of procedures, the review described robotic systems that can guide instruments along planned paths and use information from sensors to adjust their movements. Studies of robot-assisted endodontic microsurgery have reported promising accuracy in laboratory testing, and small clinical case series have provided early evidence of applicability.
“EAI provides an emerging framework in dentistry, addressing anatomical complexity through a dynamic closed-loop system of perception, planning and execution,” said the researchers.
But they added: “Rather than replacing dental professionals, EAI should be developed as a clinician-supervised, task-specific co-pilot that supports clinical planning and execution while preserving clinician judgement and accountability.”