The Stanford One Hundred Years study predicts minimal advancements of AI technology in the healthcare field in the next fifteen years. While AI functionalities will improve relative to today, machines will still serve a subservient role to human clinicians. Id like to propose an alternative future where, due to increased availability of medical information, machines will play the primary role in healthcare and human clinicians will be sparsely needed.

First, I'd like discuss the wider availability of data available for analysis, both in terms of patient reported data as well as anonymized clinician documented data. As evident by the amount of personal information freely posted on social networks, the general population is becoming numb to the idea of releasing information to unknown external entities. Letting everyone know of your status by the minute is now the norm, rather than the exception. Multiple health and fitness applications have taken advantage of this increased willingness to share by allowing individuals to share details about their daily workouts. Similarly, this increased willingness to share will allow AI technologies to analyze and detect risky behavior and promote wellbeing in day to day life. Additionally, as research initiatives such as Sync for Science take flight, patients will be given the option to donate their entire medical record for use in analytics. This step into a connected world allows machines to explore a dataset richer than we can imagine, allowing every aspect of a patients life to be examined, analyzed, and compared.

Additionally, with technologies such as the Da Vinci robot, we will have more information than ever about all types of medical procedures. While current ventures into AI technology in the surgical realm have been focused on assisting the human physician, these platforms can be repurposed to do much more. As mentioned in the Stanford study, theres already been prototypes of surgical platforms where, in addition to assisting, the platform itself observes and documents physician actions. Using such information, these smart platforms can provide much more insight into factors affecting patient outcome. Instead of relying solely on clinical documentation, these platforms will be able to correlate each action performed by surgeons with effects on patient outcome. Combined with other available datapoints, machines will be able to provide passive assistance in suggesting the best course of treatment as well as active intervention in replicating strategies with the highest impact on patient wellness.

While increased connectivity and data availability will allow AI technology to grow, perform common procedures, and suggest preventative measures, I am not implying that it will be able to fully replace human doctors. While machines may be able to conduct most, if not all, common treatments, physicians are needed for the rare procedures. By definition, a rare procedure may not be performed frequently enough for machines to learn and replicate. I imagine human physicians will be akin to specialists  the machine will take care of most, if not all, of your ailments, but will pull in a human specialist when it cannot. However, machines will be able to improve access to these human specialists  even with technology available today, the concept of telemedicine has already taken off, allowing doctors to provide treatment for patients regardless of physical location. With more sophisticated AIs and available data, telemedicine can take a big step forward  the tag team of machines to take care of common ailments and doctors to visit the unusual or uncommon patients will allow for higher throughput of sick patients and provide more comprehensive care than before.