Book Review: “The AI Revolution in Medicine: GPT-4 and Beyond”

Review by Tyler Smith, CEO, Health Data Movers.

Reading a paperback in the age of AI feels (and is) anachronistic. But until Neuralink pumps the information into our brains, such combinations of societal shattering technology breakthroughs and pre-AI workflows remain. In reading “The AI Revolution in Medicine: GPT-4 and Beyond,” it becomes evident that a combination of new technology and old processes will prevail in medicine over the next phase of technological advancement.

Given an early glimpse of GPT-4, authors Peter Lee, PhD, Carey Goldberg, and Isaac Kohane, MD penned distinctive chapters focused on their areas of expertise and the potential they have experienced or can imagine with the application of GPT-4 to medicine. They explain the basics of the incoming technology — including the fundamentals of LLM and machine learning, along with platforms like Nuance DAX — and contemplate the potential ethical pitfalls and opportunities of AI, including non-medical capabilities such as writing poetry.

Although presented as an academic text, the co-authors’ diverse backgrounds: computer scientist (Lee), physician (Kohane), and journalist / patient advocate (Goldberg) bring variety and thus life to case studies. Endowed with an introduction by the present king of AI, Sam Altman, and closed with a piece by Microsoft’s CTO, Kevin Scott, the work has all the makings of AI thought leadership star power.

Coming fresh off a viewing of “The Terminator,” the ethically positive potential of AI was a pleasant surprise, underscored by the authors’ assertion that GPT-4’s tone is more empathetic and caring than human doctors, a viewpoint reiterated recently in this post by Chris Longhurst, MD.

Quite a plot twist if the robot’s apparent empathy may force clinicians to improve their own written bedside manner. It has been awhile since I watched “Terminator 2,” but that might be a more appropriate cinematic pairing.

But as the authors take us through the modern reality of endless data capture and regulatory compliance processes (the dreaded prior auth!) that plague our healthcare providers, only a reader devoid of empathy would overlook that physicians are drowning in typing. If AI can lift the clerical burden, maybe a softer touch can find time on a physician’s daily calendar.

As an Epic implementation consultant who came of age during the Meaningful Use era, the hope presented in the book validated days and nights spent in Chronicles and Hyperspace. True the burden of typing was a byproduct of the installs performed across the country. But if the data that was gathered by such typing can now be mined and used to generate insights, and if the initial installs are seen as the first step in the shift to integrating technology into workflows (workflows that begin to substitute vocal cords for fingers), maybe it wasn’t all in vain.

A more radical prognostication of the future applications of AI in medicine would have made the book monumentally more thought-provoking. A scene similar to that contained in Lanier’s “Who Owns the Future,” wherein the reader is presented with an image of our lifestyles after physical activity disappears from the human experience (we simply live in pods) would have distinguished the work. Will AI alter our lives in such a way that medicine will change because our injuries and illnesses of the AI era will be symptoms of habits and a life we can’t even begin to imagine?

Since the authors don’t appear to aspire to thread the needle between Huxley and an academic piece, the book achieves stature as the perfect preface to an optimistic future for AI in medicine. As a fellow optimist, I’m hopeful as well, and will take such sentiment into today, where the real chapters of the story are being written.