From Code to Care, Building a Trustworthy Gateway for Applied AI in Clinical Practice
The Applied AI Accelerator (A3) Lab was established through a strategic collaboration between the University Medical Centre Groningen (UMCG) and the University of Groningen (RUG), and aims to bridge the gap between technical AI research and clinical practice. Recently launched, the lab focuses on the clinical validation of applied AI solutions within healthcare settings to ensure that they are safe, effective, and ethically sound. This partnership leverages UMCG’S secure clinical data environment with RUG’s academic rigour to transform promising research into validated medical tools. Beyond the universities, the lab acts as clinical gateway for the Dutch AI Factory (NLAIF) and is working with the Landelijke Registratie Orthopedische Interventies (LROI) to develop predictive models for joint replacements. In terms of industrial partners, the lab closely collaborates with Siemens and EPIC, along with Gleamer and PatientPlus.
The mission of the lab is to accelerate the development of AI solutions that enhance patient outcomes, improve job satisfaction of healthcare professionals, or create significant sustainable value for the healthcare system. A3 lab focuses on Explainable AI (XAI), ensuring that the rationale behind every AI-driven solution is clear, fostering trust between patients and clinicians. Furthermore, behind the lab stands an entire team dedicated to achieving this mission. Under the leadership of Academic Directors Professor Job Doornberg (RUG) and Professor Paul Jutte (UMCG), along with Lab Manager Ishaan Jagota, the A3 lab currently has 15 PhD researchers from all over the world. This diverse team is responsible for the strategic and operational excellence required to bring applied AI into the heart of healthcare.
The A3 lab focuses on three core research tracks:
- Predictive Analytics and Computer Vision. This track focuses on developing and validating tools to enable precision medicine by forecasting disease risk, diagnosis, and treatment outcomes. By integrating computer vision to interpret medicl images (X-Rays, CTs, MRIs), these tools support clinicians in personalized treatment selection and earlier interventions. This streamlined decision-making reduces cognitive load for providers and improves patient safety by minimizing procedural risks and re-admissions.
- Generative AI. In this track, the focus is on adapting and validating Large Language Models (LLMs) specifically for the Dutch healthcare system. While Generative AI has immense potential to automate documentation, existing tools often lack the language support and workflow integration required for the Dutch clinical practices. The objective is two-fold: first, to validate specialized solutions that automate repetitive tasks and generate accurate clinical notes, significantly reducing the administrative burden of clinicians. By reclaiming time for patient care, these tools aim to improve job satisfaction and reduce staff burnout, while ensuring patients receive more focused consultations. Second, the track will leverage Generative AI to empower patients, enhancing their ability to comprehend their medical information and participate in shared decision-making.
- Agentic AI. This last track explores proactive systems capable of independent decision-making and task coordination. Unlike reactive AI, these “agents” can observe environments and execute actions to support overstretched clinical teams. The emphasis is on the responsible turning and validation of these systems within complex orthopaedic workflows to ensure that they are safe, context-aware, and bias-free. The final goal is delivering trustworthy, patient-centered care that optimizes resources and enhances the entire medical care journey.
If you want to read more about the A3 lab, you can access the dedicated page on our website: https://icai.ai/lab/a3-lab/
