What is your role in the new VU Center for AI & Health?
“I’m one of the initiators of the center, along with Marleen Huysman, Robert de Jonge and Heleen Riper, and we currently make up the management team. We would very much like to bring together researchers and health experts who are interested in AI and healthcare to share experiences about the application of AI in healthcare. We would also like to create more inter- and multidisciplinary collaborations in this area on the VU Campus.”
You have been professor of Quantitative Data Analytics since 1 October 2020. Where will you be in 5 years?
“The aim of the group is to improve AI techniques, specifically in machine learning, to make it more feasible to implement them in practice. We focus mainly on applications in healthcare, but the improvements we make are, of course, more widely applicable. I hope that in five years' time, as a result of our contributions to developments in machine learning, the techniques will be more widely applicable and that we will actually see them being used by both doctors and patients.”
What is the status of your research project into colorectal cancer?
“During this project, we were able to demonstrate that by applying machine learning to a large dataset of patient records from general practitioners (completely anonymised, of course), you can create a predictive model for colorectal cancer. The predictive ability of this model exceeds that of predictive models that are based on traditional statistical methods. As an added bonus, we have also come up with new predictors—which is great, of course! The next step is to test in a clinical setting whether this can actually improve patient care.”
AI is now mainstream. How do you, as a VU researcher with a long history of working in interdisciplinary, human- and society-centered AI, view this development?
“Of course, AI isn’t the solution to every problem. I certainly think that we are still in the early stages of the application of AI in healthcare and that we are going to see an acceleration in these developments. There is a lot more data and increasing interest in AI applications in healthcare from various groups. Also, AI techniques are getting better every day. It would be best if there were realistic expectations about the contributions that AI can make to healthcare. Sometimes these expectations are enormous. I think it is important that cooperation is maintained between healthcare professionals and AI experts. Naturally, we at VU have a head start in this because of our long tradition in this field.”