"The automated screening of Mr. Smith's chest x-ray is concerning for tension pneumothorax. Would you like to view the image on your portable device?" "Of course," you respond, after noticing the paucity of peripheral vasculature in Mr. Smith's right chest. "Please order chest tube supplies to the bedside." Like many nerds who became doctors, I grew up watching movies like "Terminator," which portrayed "less-than-ideal" scenarios, where humans lost control over the evolution of artificial intelligence (AI). As we enter into this exciting phase of medical technology integrating with AI systems, physicians will need to guide the way forward. Not so long ago, the electronic medical record was bestowed upon us, transforming medicine in many positive ways, but in several ways that were "less-than-ideal." We are again at a pivotal crossroads, and our leadership is needed now more than ever. Our roles and capabilities as physicians are likely to be greatly augmented by AI, rather than deflated, if we become involved actively in its development.
My goal for this talk, as a physician and computer scientist, is to provide the audience with a framework to understand machine learning, how it works, its current limitations, and how it will transform medicine. We will share several exciting examples of how machine learning is already shaping health care and data science, and how physicians can provide engineers with context for its implementation. After this talk, you will leave with some fresh ideas, some working language for dialogue, and most importantly, a way to make yourself useful to our future robot overlords. Whether you are about to be a resident, an attending who may feel out of touch with "this tech stuff," or a robot, you should come to hear about AI in medicine, which is closer to your reality than you might think.