Within 10 years time, there will be enough people to staff the health care system. It is hoped that AI technology - robots, smart technology, eHealth, smart apps, etcetera - may contribute to solve this staff shortage. The main areas where AI technology shows promise are support for the elderly when their mental and physical abilities decrease, and support for patients with chronic diseases or disability. But it is also hoped that smart technology can also play a role in prevention, by helping all people, healthy or not, to adopt and maintain a healthier lifestyle.

To ensure that AI’s promises can materialise, it will be essential to take the needs, abilities and attitudes of the intended users into account. Not only is it imperative that users be involved in the early development phases of new technology, it is essential that we know the user before even starting to develop new technology.

Although everyone pays lip service to the idea of user-centered design, in practice, most AI technology is mainly developed in AI Departments or companies by technologists. I will argue that health technology can only be successfully introduced when it is designed within truly multi-disciplinary teams. I will provide a few examples from my own research and that of others to show that knowledge about the intended patients and users is imperative for AI technology in health care to ever become successful.

Fokie Cnossen was trained as an experimental and neurobiological psychologist at the University of Groningen, and worked as a human factors researcher in traffic psychology. She now is lecturer at the Institute of Artificial Intelligence and Cognitive Engineering at the University of Groningen, where she is head of the Cognitive Engineering group.

Her research interest lies in optimising human performance, either by designing better technology (i.e., cognitive ergonomics and human factors research) or by improving education (i.e., more fundamental research on skill acquisition). Her research is mainly focussed within the medical and health care domain.