2015 : WHAT DO YOU THINK ABOUT MACHINES THAT THINK?

eric_topol's picture
Professor of Genomics, The Scripps Translational Science Institute; Author, The Patient Will See You Now
A New Wisdom of the Body

 

Back in 1932, Walter Cannon published a landmark work on human physiology—The Wisdom of the Body. He described the tight regulation of many of our body's parameters, such as hydration, blood glucose, sodium, and temperature. This concept of homeostasis, or auto-regulation, is an extraordinary means by which we stay healthy. Indeed, it represents a machine like quality, that our body can so finely tune such important functions.

Although it has taken the better part of a century, we are now ready for the next version—Cannon 2.0. While some have expressed marked trepidation about the rise of artificial intelligence, this capability will have an extraordinary impact on preserving our health. We are quickly moving to "all-cyborg" status, surgically connected to our smartphones. While they have been called prosthetic brains, "smart" phones today are just a nascent precursor to where we are headed. Very soon the wearable sensors, whether they are Band-Aids, watches, or necklaces, will be accurately measuring our essential medical metrics. Not just one-off assessments, but continuous, real-time streaming. Obtaining data that we never had before.

Beyond our body's vital signs (blood pressure, heart rhythm, oxygen concentration in the blood, temperature, breathing rate), there will be quantitation of mood and stress via tone and inflection of voice, galvanic skin response and heart rate variability, facial expression recognition, and tracking of our movement and communication. Throw in the analytes from our breath, sweat, tears, and excrements into the mix. Yet another layer of information captured will include our environmental exposures, ranging from air quality to pesticides in foods.

None of us—or our bodies—are smart enough to be able to integrate and process all of this information about ourselves. That's the job for deep learning, with algorithms that provide feedback loops to us via our mobile devices. What we're talking about does not exist today. It hasn't yet been developed, but it will. And it will be providing what heretofore was unobtainable, multi-scale information about ourselves and—for the first time—the real ability to pre-empt disease.

Almost any medical condition with an acute episode—like an asthma attack, seizure, autoimmune attack, stroke, heart attack—will be potentially predictable in the future with artificial intelligence and the Internet of all medical things. There's already a wristband that can predict when a seizure is imminent, and that can be seen as a rudimentary, first step. In the not so distant future, you'll be getting a text message or voice notification that tells you precisely what you need to prevent a serious medical problem. When that time comes, those who fear AI may suddenly embrace it. When we can put together big data for an individual with the requisite contextual computing and analytics, we've got a recipe for machine-mediated medical wisdom.