UMCG develops national framework to bring trustworthy generative AI into clinical practice

News
Would you trust an AI assistant to write your medical records?
Job Doornberg en Bart Scheerder

Generative AI – such as voice-enabled documentation that automatically turns conversations into written reports – holds great promise for reducing the administrative burden in healthcare. Currently, healthcare professionals spend up to 40% of their time on paperwork – time that could be better spent with patients. But before this technology can be widely adopted, we need to be certain it is safe, effective and acceptable in practice.

That’s why UMCG is taking part in the new RIGH:T consortium, a collaboration of nine healthcare and knowledge institutions working to ensure the responsible use of generative AI in healthcare. Together, they are developing a validation framework: a practical assessment tool that will help healthcare providers and institutions determine whether an AI application meets the necessary standards for safety, quality, and usability in clinical settings. It also offers guidance to developers and suppliers aiming to bring responsible AI solutions to market.

More than just the technology

The RIGH:T consortium goes beyond testing technical accuracy. It also examines how AI impacts clinical practice, cost-effectiveness, professional and patient acceptance, and broader societal issues – including sustainability and ethics. Once the validation framework is in place, the consortium will continue working on long-term implementation and scaling across the healthcare sector.

A national effort, driven by healthcare itself

What makes RIGH:T unique is that it was initiated from within the healthcare field. It brings together organisations committed to sharing knowledge and experience: Elisabeth-Tweesteden Hospital, UMCG, Erasmus MC, UMC Utrecht, LUMC, Maasstad Hospital, Frisius MC, TU Delft and GGZ Noord-Holland-Noord. By working together, they aim to avoid duplicating efforts and develop a unified approach to this rapidly evolving technology.

Despite the growing interest in AI, less than 2% of AI models are currently in actual clinical use. This is largely due to a lack of validated applications, clear guidelines, and slow adoption. The RIGH:T consortium aims to change that – accelerating the use of generative AI in healthcare, while keeping safety, responsibility and quality at the forefront.