Depending on your perspective, Artificial Intelligence (AI) either brings thoughts of a future enhanced by technology or strikes fear in your heart that robots will take over the world. Some of the brightest minds in technology (I’m talking about you Elon Musk) have warned humanity about the dangers on the horizon, yet it’s breathtaking to envision just how much AI can improve our lives. I for one welcome our new robot overlords.
In Deep Medicine, cardiologist Eric Topol takes readers on a fantastic journey to a not-too-distant future where AI works side-by-side with the medical community to improve all aspects of health care, from individualized medicine to improved productivity and lower costs to virtual personal wellness assistants (think Amazon Alexa on steroids). Topol’s research points to numerous advances either currently underway or in development and paints a picture of a true partnership that is already shaping the health care environment.
I choose to take an optimistic view of this future, perhaps because I admire science and what it has done and can do for humanity. When I bring up the future of medicine in conversations, I often see fear come across people’s faces rather than wonder. It happened today at a luncheon when the discussion turned to the future of health care and someone at the table brought up futurist Ray Kurzweil’s belief that in the coming years, we’ll all have nanobots injected into our bloodstream designed to hunt down and kill infections. A woman at the table looked downright mortified, and her look became even more uncomfortable when I followed by describing Kurzweil’s conviction that in the next couple of decades mankind will reach a singularity where man will essentially merge with machines. For some this concept means the end of homo sapiens, but for me and other cheerleaders for AI and machine learning it simply means the beginning of our next phase of evolution. Who’s to say evolution isn’t meant to include the merging of flesh and technology?
I understand why AI might frighten some people. And I agree mankind must move forward with a set of ethical guidelines to ensure we don’t end up in some dystopian brave new world. But I think a future with AI is brighter than a future without it. For me, science and technology represent progress. Just 100 years ago, human life expectancy was around 50 years. Today, thanks in large part to science and technology our life expectancy is around 80. Many experts believe that a person born in 2100 will have an average life expectancy of 100 meaning many of us will live well beyond the century mark. I can’t help you if you don’t want to live that long, but I’d suggest your worldview may be skewed by the fact that you equate old to unhealthy. Topol’s point is that AI and machine learning will improve not just our longevity but perhaps more importantly our quality of life.
The book, however, does not fail to alert readers to the potential pitfalls of a health care system reliant on AI. Perhaps the biggest issue for Topol, and rightfully so, is the concern over privacy. If bots and databases have access to our health information – including our genetic code – doesn’t that set up a scenario whereby our private health data can be stolen or even sold to the highest bidder? Or might we end up in a place where our employer, or the government, can deny certain rights or privileges because of our health? I’m no Pollyanna. I can imagine a situation where a person could be denied employment or insurance coverage because they have a health issue. However, Topol argues that we need to set up safeguards for just these kinds of things and if we do the pros will far outweigh the cons. And not for nothing, it’s not as if the American health care system is working now. Even though America pays thousands more per capita for its citizen’s healthcare we rank at or near the bottom of most global ratings systems in terms of measures like quality and efficiency. We can only go up from here, and technology will play a key role.
Topol also raises concerns over where technology will reside within the healthcare system. By that he means if we let AI replace doctors, we may gain some efficiencies and diagnosis improvements, but we will lose the patient/doctor connection or “presence” that he and others believe is paramount to quality care. Rather, he argues that the best future sees AI working alongside medical practitioners in order to free them up to focus on patient care. Doctors today spend more time entering data into electronic health records via keyboard or recorder and less time listening to and examining patients. If AI is able to manage the routine parts of healthcare and assist doctors with diagnosis using machine learning and big data that will give doctors more time to interact with patients, something all medical professionals agree equates to better results. If you’ve ever sat in a doctor’s office while he or she typed notes into a tablet or laptop instead of looking you in the eyes you know what I’m talking about. These days it’s not unusual for a doctor to spend less than 10 minutes with a patient!
“The greatest opportunity offered by AI is not reducing errors or workloads, or even curing cancer: it is the opportunity to restore the precious and time-honored connection and trust—the human touch—between patients and doctors,” Topol writes in Deep Medicine.
That said, there are so many areas of medicine where AI can help us make giant leaps in diagnosis and treatment. AI has remarkable advancement opportunities in areas like pathology and radiology where machines can look at millions or billions of inputs to see patterns that the human eyes could never see. It can identify correlations that might never be noticed by human researchers. It can scan through billions of pages of studies and articles and identify a diagnosis in a fraction of a second, where humans could never consume or recall the vast knowledge base that medicine relies upon. It can compare billions of patient records and uncover connections that can lead to new, more effective, treatment protocols. It can compare genetic data with pharmaceutical data and recommend the exact right drug to treat an individual’s specific medical issue.
One area that I found fascinating (and really exciting) is the potential for virtual medical assistants that can work with patients in the clinic or at home to diagnose and suggest treatment. If you have a digital assistant at home (like an Amazon Alexa or Google Home) or a smart watch or smart phone you’ve already seen how this works. By 2018, more than 60 million Americans were using these AI-powered devices and that number will grow. It’s projected that by 2020, 75 percent of American homes will have at least one voice-activated personal assistant. Indeed, we have a Google Home device in every room of our home and access Google Assistant on our smartphones multiple times per day. And while medical applications for these devices today are pretty much limited to counting steps or providing information from WebMD, many people (including myself) now have a smart device that can check one’s pulse rate and some even check blood pressure. The medical applications for these devices are increasing rapidly.
“Offering a glimpse at how wearable technology may help flag potential health problems, the Apple Watch was able to detect atrial fibrillation (AFib) in a small group of people who had received an alert of an irregular heartbeat, said researchers presenting results from the Apple Heart Study.” – American College of Cardiology, 2019
The future will include glucose monitoring for diabetics, gut biome assessment, sleep improvement, stress reduction, and more. Topol lays out a host of amazing ideas, some that are already in development. He offers several potential scenarios where an individual can begin a dialogue with a home-based AI that can results in immediate diagnosis via voice input and/or device input (one even includes a scenario where the person is asked to place his phone up against his stomach where he is having pain and the device uses ultrasound or some mechanism to determine that the patient has gallstones and should see his doctor for treatment). The future is only limited by the imagination of data scientists and inventors who can design and build all kinds of input devices.
Again, Topol is in no way suggesting AI should replace doctor care. But if AI can diagnose many common and treatable issues at home the time and cost savings will be dramatic. I’ve already had a doctor diagnose an infection by looking down my throat via video conference on my mobile phone – it’s a simple step to having that same diagnosis come from a machine without the need for human interaction.
Frankly, all of this excites me and makes me hopeful for the future. And Topol didn’t even get in to other medical advancements in this book, such as precision medicine, 3D printed spare organs, or CRISPR gene editing. I admit that part of the reason for my interest in these areas is because I have a damaged heart from a heart attack I experienced in 2011. The idea that my heart can be returned to normal through new medicines, gene editing or stem cells is, well, heartwarming. This technology came too late to keep me from having the heart attack in the first place, but it’s not too late to both improve my life and perhaps extend my life. And while I’m not exactly holding out for the singularity, I wouldn’t mind living a long and healthy rest of my life.