As AI reshapes clinical workflows faster than regulatory and ethical frameworks can keep pace, one surgical specialty offers a unusually detailed window into both the promise and the structural risks of algorithmic medicine. Hand surgery — with its blend of fine motor precision, imaging complexity, and rehabilitation demands — has become an unexpected proving ground for clinical AI across the full care continuum.
A comprehensive review in the Journal of Hand Surgery, European Volume maps AI's current footprint across fracture detection (distal radius and scaphoid), osteoporosis estimation from plain radiographs, triangular fibrocartilage complex identification on MRI, carpal tunnel syndrome diagnosis via neurophysiological and ultrasound-based models, and bone segmentation for dynamic wrist analysis. On the procedural side, large language models are being applied to surgical triage and coding, while computer-vision systems perform phase and gesture recognition from intraoperative video. Autonomous microsurgical prototypes and telemanipulator platforms for supermicrosurgery represent the frontier. Postoperatively, remote photoplethysmography and video-based mobility tracking are entering rehabilitation monitoring, and multimodal sensing is advancing myoelectric prosthesis control. Prognostic models for carpal tunnel release and thumb carpometacarpal osteoarthritis show mixed performance — an honest acknowledgment rarely foregrounded in AI enthusiasm cycles.
What distinguishes this review is its systematic attention to risk taxonomy: algorithmic bias is parsed into data, transposition, normative, and annotation subtypes — a granularity absent from most clinical AI commentary. The concern about automation bias and skill erosion deserves particular weight. Surgeons trained with AI-assisted navigation may develop procedural dependency that degrades independent capability, a phenomenon already documented in aviation. This review is largely descriptive and stops short of quantifying effect sizes or comparative performance benchmarks across tools, limiting its actionability. Still, it is a valuable orientation document for clinicians and institutions navigating AI adoption decisions in surgical specialties.