Our new paper in npj Digital Medicine introduces RAVEN, an AI second marker that combines VR video, action logs and examiner assessments to support review of clinical exam marking.
In a retrospective evaluation involving 120 students and 442 ratings, a confidence-gated combination improved agreement with an examiner panel, with the largest gain in borderline-fail cases. The system is intended to support human review of assessment decisions.
What helps people and AI work well together in healthcare? Our review in npj Digital Medicine brings together 140 studies to map how collaboration is evaluated and what shapes its success.
The review highlights task fit, workflow integration, training and trust, alongside gaps in evaluating patient outcomes, accountability and safety. Our new explainer walks through the evidence and its implications.