My not-yet-expert* response whenever my classmates joke about me building some AI robot-doc to replace them: "AI will become just another tool in a physician's

My not-yet-expert* response whenever my classmates joke about me building some AI robot-doc to replace them: "AI will become just another tool in a physician's toolkit." AI will not replace the role of a doctor. However, the role of a doctor might shift to overseeing the AI in very specific scenarios. Furthermore, I'll add that there's a lot more to medicine than making diagnoses. As it stands right now we have no limit of promising applications of AI in healthcare. The bigger issue is getting these applications into the clinic. Very few of these boundless applications are actually implemented in a real-life setting that can affect patient outcomes. I can't tell you how many studies are published proclaiming, "HEY AI CAN DO X BETTER THAN DOCS." Cool study, bro. Now can you actually get that into clinic and start saving lives? Suchi Saria at JHU is one of the few people I know that has bridged that gap. Other resources for those that are interested: - http://mucmd.org/ http://mucmd.org/ - Baxt, William G. "Application of artificial neural networks to clinical medicine." The Lancet 346.8983 (1995): 1135-1138. *Programmer MD student about to start my PhD in comp sci specifically machine learning + healthcare.