ABSTRACT
Current technological and medical advances lend substantial momentum to efforts to attain new medical certainties. Artificial Intelligence can enable unprecedented precision and capabilities in forecasting the health conditions of individuals. But, as we lay out, this novel access to medical information threatens to exacerbate adverse selection in the health insurance market. We conduct an interdisciplinary conceptual analysis to study how this risk might be averted, considering legal, ethical, and economic angles. We ask whether it is viable and effective to ban or limit AI and its medical use as well as to limit medical certainties and find that neither of these limitation-based approaches provides an entirely sufficient resolution. Hence, we argue that this challenge must not be neglected in future discussions regarding medical applications of AI forecasting, that it should be addressed on a structural level and we encourage further research on the topic.
