Close-up on a monitor at the hospital monitoring a patient’s vitals – healthcare and medicine … [+]
One of the areas where new artificial intelligence tools are doing most is in the world of medicine.
Partly that’s because this is a data-driven enterprise – big data has the capability to help heal disease, discover new drugs, and even work on the human genome itself.
So this kind of work started prior to the AI revolution, but AI supercharges what scientists are able to do.
Looking back through what we’ve reported on this year, these are some of the biggest stories around what we’ve been able to do with our new technologies.
Artificial Intelligence and The Cartography of Medicine
One of our panel discussions from earlier this year talks about how to take a linear view of data and make it higher-dimensional, to see the cartography of things like disease and treatment and apply that to clinical practice. There are also some notes about genomics and other research in here, and thoughts on how to innovate pharmaceutical work.
This set of ideas has a lot of application to modern medicine, so if you’re interested, check it out. You can also watch the video.
Working Through the Puzzles
What about when you have contradictory or challenging medical information to work with?
This past summer, we reported on the MIT Media Lab’s Abhishek Singh talking about vitamin E supplements and their actual outcomes in the human body. He used the analogy of the early explorer moving through uncharted waters – how do you find what’s real, and what’s true?
Scientific 3D reconstruction of a spanish galleon armada led by fernando magellan in the 16th … [+]
Part of the process involves what Singh calls the “self-aggregation of data,” from the patient perspective, and different kinds of new mapping of healthcare processes.
AI and Heart Health
We know that cardiovascular disease is an epidemic, but in a presentation that we covered, Collin Stultz suggests that AI can have a good preventative effect by globalizing the data around heart-related health. The heart is a complex organ and muscle, and there seems to be a lot of potential for deep learning to help there.
Equity in Healthcare: Getting Rid of AI Bias
Then there’s this struggle to make healthcare fairer and more equitable for citizens.
This is an example of collaborative AI, where Marzyeh Ghassemi suggests that the AI systems can work alongside doctors to figure out where bias lies, and how to remediate it.
This also highlights that fundamental idea that most of our AI systems should be assistive and not automated, that they should be decision support tools, rather than making their own decisions devoid of human intervention.
More Innovation?
I wanted to end with these companies that are doing a lot in the AI world and having a big effect on healthcare.
There’s GNS, where DNA research is contributing to improving outcomes.
There’s OpenEvidence, a company that is trying to evolve that process of getting all the data where it’s needed in order to do the most good — again, we started with this in the pre-AI era, with the electronic health records and The HITECH Act and other work at the HHS, but now private sector partners are helping to move the ball forward.
And there’s StabilityAI, which addresses one of the biggest cases of AI in healthcare, which is radiology.
Basically speaking, AI has been demonstrated as valuable in helping to read radiology results — x-rays, CT scans, MRIs, etc. Because we use these types of radiology so commonly and so broadly, we get a lot of benefit out of applying AI to these processes. And again, this can be assistive, where doctors get a helping hand and become more productive and more accurate in looking at what they’re seeing in radiology results.
That’s a little bit of a survey of where we’ve been this year in the healthcare world, and what our experts are reporting when it comes to giving us new tools to address health crises, practice better preventative medicine, and to modernize healthcare around the world.
