Embodied artificial intelligence (AI) must be tested in the clinical environments where it will operate, but building realistic, robot-testable settings is costly and difficult to scale. Here we show that routine clinic images can be transformed into operational digital twins for task-based evaluation of embodied AI. Using 39 ophthalmic clinic scenes, we converted single photographs into editable, simulator-ready environments and assessed reconstruction quality, room-scale geometry, mesh grounding, multi-robot feasibility, perturbation sensitivity and closed-loop policy performance. The reconstructed scenes preserved workspace structure, while local editing enabled controlled device reconfiguration. Device meshes, collision proxies and semantic anchors converted visual reconstructions into contact-aware simulation scenes. Across three robot embodiments, shared task targets showed different patterns of reachability and contact feasibility. Small device translations and rotations produced task-specific changes in contact margins that were not captured by visual similarity alone. Digital-twin trajectories also supported local policy learning and closed-loop evaluation. These findings establish operational validity as a key principle for clinical digital twins and provide an intermediate layer between offline development and physical deployment of embodied AI in healthcare.
Xinyuan Wu, Jingrao Zhang, Meng-Di Xu et al.· 0 citations
Robotic systems are entering clinical care, but full autonomy remains constrained by patient variability, safety requirements, dynamic clinical environments, and the need for professional oversight. Human–robot collaboration (HRC) offers a more realistic path: clinicians retain judgment and responsibility, while robots and AI support sensing, planning, action, and decision‐making within defined task boundaries. In this narrative review, we examine recent advances in clinical HRC from a systems perspective, covering system architecture, learning and control, safety mechanisms, autonomy assessment, and real‐world deployment. We propose a dual‐brain framework, comprising a professional brain for clinical reasoning and decision support, and a physical brain for embodied sensing, planning, and task execution. Using examples from imaging, rehabilitation, surgery, and outpatient care, we argue that clinical autonomy is developing in stages from clinician‐supervised imaging and decision support to bounded robotic subtasks and workflow‐level coordination, instead of through unsupervised replacement of healthcare professionals. We further discuss how robustness, trust, accountability, and human factors shape safe adoption in practice. By framing autonomy as collaborative rather than substitutive, this review provides a unifying foundation for designing and evaluating the next generation of autonomous clinical systems.
