Popular media like to report that artificial intelligence (AI) might soon replace human clinicians, though reality is that we are far from ready to broadly implement such algorithms in clinical practice. Important barriers remain that prevent successful human-AI collaboration, such our limited understanding of how clinical reasoning can best be supported by AI and a lack of trust in artificial intelligence. With this project, we will set out to identify minimal necessary characteristics to 1) support and improve clinicians’ diagnostic process and 2) establish trust in AI, which will lead to the development of recommendations for optimal conditions for human-AI collaboration.