Operationalising Inclusion for Participatory Design: Worked Examples for TRIPOD+AI & PROBAST+AI

Artificial intelligence (AI) and machine learning (ML) are rapidly transforming the healthcare landscape, enhancing diagnostic accuracy, personalising treatments, and improving operational efficiency. However, alongside these advancements comes a critical concern: the potential for AI models to perpetuate, or even exacerbate, health inequities. This paper examines the complex potential of AI in health, and how it can both improve and undermine equity. The key to addressing these challenges lies not only in developing advanced algorithms but in fostering responsible and inclusive development. We argue that to operationalise inclusion in AI systems, research co-design should be prioritised in order to integrate the perspectives of diverse knowledge users, including patients, clinicians, and community partners, shown through worked examples in TRIPOD+AI and PROBAST+AI. This approach encourages future work to rethink the role of knowledge users in the development of AI for healthcare.