AI apparently detected a cancer pattern in a dataset that humans had missed for years, so why are clinicians still so resistant?

I work in oncology. I follow the research. There are now multiple published studies showing AI systems identifying patterns in imaging and pathology data that experienced clinicians missed or identified later. The AlphaFold work on protein structures alone has changed the speed of drug discovery research fundamentally.

So I find the resistance among some of my senior colleagues genuinely puzzling. These are not technophobes. They are intelligent, evidence-based practitioners who are resistant specifically to AI tools in clinical settings in a way they are not resistant to other new evidence.

I have my own theories about why but I want to hear from people who have studied this or experienced it from the clinician side. What is actually driving the resistance and is it rational given the evidence?