In the short term, it will be diagnostics. Entering in symptoms and lab results into a machine learning algorithm can help a doctor eliminate conditions or possibly add to the set of illnesses to be considered. Down the road, it will be in research and determining the effectiveness of various treatments based upon the patient's factors. But in the short term, the biggest bang for the buck will be in diagnostics.

Agree its hard to see much in short term versus long term. It seemed clear from JP Morgan reports that investors and market see more functional or time-saving aspects from AI technology applications short-term. The biggest opportunity is in true AI making care decisions after working many years through developing the intelligence that could make that happen--resulting in total care cost efficiencies not possible today.

As a full-time clinician , I see AI (Augmented Intelligence) and Deep Learning platforms as having a secure role in the future of Healthcare. The system now lacks the fluidity and logic that AI could bring to it. From improving access to care, triage of patients, selection of dx tests, interpretation of results and data (especially imaging data!) selection of treatments to halting of therapy when futility is reached ( Palliative Care ) ...the possibilities are indeed endless.
I don’t see it as a substitution or replacement of providers ( although the discipline will have to evolve to adapt to AI availability and results. and yes, some providers will be displaced!), but as an amazing adjunct, additional resource to make Healthcare what it should be in today’s age. Let “computers” do what they do better... let humans do what we do better ( empathy, compassion, intuition, HUMANITY).
Leting Doctors (providers) be better HUMAN caregivers. Maybe allowing no more time with patients, but BETTER quality time with patients.
RafaelGrossmann.com

@ZGJR

Indeed diagnostic as you all mentioned, then virtual health assistant supporting patient to stay in track with therapy. Advancement in treatments by analyzing data, as an example reviewing big data to support with the creation of medicinal. AI could save costs and time (e.g. assisting with repetitive jobs) thanks to an higher efficiency and precision.

I see lot of short term use cases in "computer-assisted diagnostics". And this is becoming more evident in developing countries where manual role is quiet high in diagnostics. Another area where lot of AI is talked about is Medical Second Opinion. This area is mostly driven by analyzing similar conditions in history and providing benchmark for treatment. I have also got request for Tele-Medicine applications for primary healthcare i.e.; applications like skin disease detection, vitals analysis etc. Healthcare needs are outpacing the man-power creation in the domain, so definitely machine need to come for rescue both short term and long term.

The cost savings that AI could potentially bring healthcare is breathtaking:

1. The amount of data that can be processed over the next ten years could revolutionized everything we know about the human body and subsequently improve diagnosis and treatment plans with less physician interaction than is required today.
2. Increased certainty in diagnosis because the amount of data an AI system could draw off of for information and processing would exceed the capabilities of our best teams of physicians in a collaborative effort.
3. Reduced error rates in invasive procedures as they become more automated.
4. AI could drastically decrease wait times for appointments and procedures because specialization can be more easily duplicated in that world than it can in us humans.
5. The need for a specialized, human controlled environment has kept healthcare research and patient treatment from becoming streamlined in the ways that AI can eventually produce.
6. Healthcare payment determination and processing!! The amount of human investment into the qualification and payment process costs more than the treatment in almost all common healthcare interactions. The hours of labor that could be saved is massive. Also, the reduction rates processing could be drastically improved.

Time & resources management.

Healthcare is so complex as an ecosystem that AI and Machine learning could help.
We all can understand why for example Hospitals' administrators are regulating expenses.
Now what is mission critical and which resources should be allocated, or not to it.
So many external companies hired at great costs are pouring their information from survey or theoretical models that in the end do not fit with the local context.
AI & ML are founding their reasoning and reporting based on hard facts & data.

If you apply Asimov's laws of robotics, not for a entertainment sci-fi purpose, but for its own purpose...there is no room for human interpretation or politics debates in favor of one or another community, lobby or circles (of political) influence.

It is my humble opinion.

I Mostly agree with what my peers expressed here, from a public audience standpoint and not from a physician's specialist one.

Regards,

Stephan J. Metral IoT innovator
V-P of Digital Transformation at XD Innovation Canada