Embodied Artificial Intelligence in Healthcare: Foundations, Applications and Future Challenges

This study aims to systematically review the current applications, technological foundations, and deep-seated challenges of Embodied Artificial Intelligence (EmAI) in healthcare. It validates EmAI’s immense potential to drive the transformation of medical intelligence from “screen-based intelligence” to “scenario-based intelligence,” which is crucial for building next-generation smart healthcare systems. The methodology involves a systematic review focusing on summarizing EmAI’s performance in clinical interventions and assisted living, core technologies e.g. multimodal perception, large-model reasoning, and world models, as well as the urgent need for federated learning and responsible research and innovation principles. Quantitative findings confirm EmAI has evolved into an intelligent collaborator rather than a mere auxiliary tool: AI-assisted surgery improved surgical precision by 40% while significantly reducing complications and operating time. Within assistive technologies, EmAI substantially empowered the independence of visually impaired individuals, embodying a “human-centered” design philosophy. These conclusions support the core argument that EmAI is a key driver of healthcare transformation. Future research should focus on advancing FL to address data generalization and enhance the real-time decision-making robustness of world models. Prioritizing the establishment of clear ethical, legal, and fairness frameworks is essential to ensure EmAI evolves into a safe, reliable, and universally accessible force for healthcare transformation.