The One Skill That Could Save Medicine
How AI Might Finally Make Care Feel Human Again
Imagine waking up one morning to a diagnosis that changes everything.
Heart disease. Cancer. A chronic condition that suddenly makes every day feel borrowed. Fear rushes in—questions about treatment, survival, quality of life. You look to the doctor for answers, but the appointment is rushed, the notes endless, the system overwhelmed. The human connection you crave gets lost in paperwork and prior authorizations.
We chase better scans, new drugs, cutting-edge therapies, yet we overlook something fundamental: the system itself is failing us. Clinicians burn out from administrative burdens. Patients wait weeks for care that feels impersonal. Technology has promised relief for decades—electronic records, decision support—but often delivered more frustration than freedom.
What if the missing piece isn’t another gadget or protocol? What if it’s finally using technology to restore what medicine has lost: time for presence, space for empathy, and trust in the process?
This question has haunted me since my own brushes with broken systems—as a veteran navigating VA care, watching my sister’s pancreatic cancer journey, seeing how administrative sludge steals moments that matter.
It came into sharp focus in my recent Passion Struck conversation with Dr. Robert Wachter, whose new book A Giant Leap: How AI Is Transforming Healthcare—and What That Means for Our Future doesn’t sell AI as a miracle. It asks the harder question:
Can AI succeed where every previous digital wave failed—by making care better, not just faster?
A Warning Shot in the Exam Room
Imagine a quiet exam room, the soft hum of fluorescent lights, and a patient seated, a familiar knot of uncertainty in their chest. A doctor appears on the screen: measured voice, gentle eye contact, unhurried cadence. They explain the MRI findings with clarity and warmth—acknowledging fear without amplifying it, offering next steps without false reassurance. The patient exhales, nods, feels seen.
Then the real physician steps into frame from behind the monitor. The illusion dissolves in an instant. What felt intimate was synthetic. What felt human was engineered.
This is no speculative demo; it is a live demonstration from the Mayo Clinic, where digital twins of clinicians now deliver preliminary explanations, scale scarce expertise to rural clinics and overburdened urban practices, and bridge gaps in access that have persisted for generations. The technology is breathtaking in its precision: trained on vast corpora of medical dialogue, it reproduces not just facts but tone, pause, empathy—qualities once thought to be the irreducible core of the physician’s art.
And yet the moment carries a shadow.
If a machine can simulate compassion so convincingly that a frightened patient cannot immediately distinguish it from the real thing, what becomes of trust?
If deepfake avatars can be spun from the same underlying models—speaking with a clinician’s face and voice to deliver misinformation, coercion, or outright harm—what safeguards remain between authenticity and deception?
The scene is not merely a technological milestone; it is a moral threshold. It compels us to ask the question that will define the next era of medicine:
When artificial systems can convincingly perform the relational labor of care, what is left that is uniquely, irreplaceably human?
Gradually, Then Suddenly
Healthcare has clung to analog relics longer than nearly any other sector, a quiet testament to how deeply relational and regulated care truly is. In 2026, fax machines remain the backbone of much clinical communication—accounting for 70–90% of exchanges in many settings, including referrals, lab results, and insurance documentation.
Faxing persists because it offers a perceived reliability and security that digital alternatives have yet to fully supplant, even as cloud-based options slowly migrate in. Pagers (or beepers), once thought extinct, still hum in the pockets of clinicians in hospitals and emergency departments—valued for their durability, long battery life, and one-way urgency in high-noise environments where phones might fail.
This inertia is the residue of a system built on caution, liability, and human stakes.
For decades, we’ve had the “gradually”: electronic health records (EHRs) rolled out with grand promises of liberation, only to deliver alert fatigue, documentation burdens, and pervasive burnout.
Recent studies show physician burnout rates hovering at 45–50% in the U.S. (down slightly from pandemic peaks but still alarmingly high), with EHR usability and notification overload cited as major drivers—clinicians often spend more time on screens than with patients, eroding the very presence that defines healing.
Now, at the breaking point, AI arrives not as another incremental tool, but as a potential rupture. Clinician shortages intensify (with projections of hundreds of thousands of unfilled roles by the end of the decade), access crises deepen (wait times stretching months in many regions), and costs spiral (healthcare nearing 20% of U.S. GDP).
Generative AI changes the equation: it parses unstructured notes, interprets spoken language, synthesizes vast datasets, and generates responses that feel remarkably human. This convergence—technological maturity meeting systemic desperation—creates the “suddenly.” Not overnight magic, but a tipping point where the accumulated weight of inefficiency finally demands something more radical.
The question isn’t whether AI will arrive—it’s already here in scribes, triage tools, and diagnostic aids.
The question is whether it will finally resolve the contradictions that have plagued healthcare for generations: freeing clinicians to be present again, empowering patients without overwhelming them, and rebuilding trust in a system that has quietly eroded it.
The Tools That Could Heal—or Harm
Picture a primary care physician in 2026: twenty-five patients scheduled, charts layered with years of unstructured notes, inbox overflowing with messages, prior authorizations pending, billing codes demanding justification. For too long, half the day dissolved into documentation—typing, scrolling, clicking through endless dropdowns—leaving scant room for the listening, diagnosing, and comforting that drew most to medicine.
Now, ambient AI scribes sit quietly in the room (or join via app), capturing dialogue in real time, distilling it into structured notes with near-perfect fidelity, surfacing relevant history, and drafting thoughtful patient replies or discharge summaries.
Inbox triage ranks messages by urgency, suggests evidence-backed responses, and auto-generates prior-authorization appeals aligned with guidelines. Diagnostic copilots scan lab and imaging results, flag subtle anomalies, and propose ranked differentials. In radiology, models read scans faster and—in many contexts—with greater consistency than fatigued humans; in operating rooms, they aid pre-operative planning and real-time guidance.
These are not distant prototypes. They are deployed across systems from Mayo Clinic to Cleveland Clinic to community practices, frequently reclaiming 20–40% of administrative time. Recent studies show modest but meaningful drops in after-hours “pajama time,” reduced burnout metrics, and patients who report feeling more heard because the clinician is finally looking at them, not a screen.
Yet liberation carries latent costs.
The same model that crafts elegant notes can hallucinate a fabricated allergy or interaction with serene authority. Cognitive offloading risks dulling situational awareness: when the tool proposes a plan, acceptance may outpace independent reasoning. Biases in training data can quietly widen disparities—underdiagnosing in underrepresented groups or skewing recommendations. For patients turning to consumer-facing chat tools, a confident but incorrect reassurance can delay care, fracture trust when the mistake emerges, or cause harm.
The tools are powerful. Without deliberate guardrails, power becomes peril. The decisive factor is not algorithmic brilliance, but intentional design: where—and how firmly—we keep humans in the decision loop.
The Quiet Hope in the Middle
This moment isn’t about replacing humans—it’s about redesigning roles so care can feel human again.
Imagine a future where AI has quietly taken over the soul-crushing paperwork: scribes that free clinicians from screens, triage that surfaces only what truly needs attention, and copilots that flag risks without dictating decisions. The physician sits across from the patient—not glancing at a monitor every few seconds, but truly present.
Eyes meet. Stories are heard. Bad news is delivered with the weight only a fellow human can carry. Burnout eases not because the work vanishes, but because the dehumanizing friction does—leaving room for the soul of medicine: judgment in uncertainty, empathy that can’t be scripted, moral responsibility that no algorithm can shoulder.
Patients, too, step forward as true partners. Armed with AI-synthesized insights from their own records, they arrive prepared with thoughtful questions, no longer passive recipients but co-navigators of their care. Trust rebuilds not through perfection, but through transparency: explainable tools, human oversight where it matters most, incentives aligned with outcomes rather than volume.
It echoes my own path. For years, I suppressed trauma—PTSD, traumatic brain injuries—treating symptoms in isolation, piecemeal fixes that never reached the root. Only when I addressed the whole person—mind, body, history—did real healing begin. AI holds the same potential for medicine: to augment rather than sideline, to restore wholeness rather than fragment it—if we choose wisely. If we insist on keeping humans in the loop. If we design not just for efficiency, but for dignity.
Listen to the full conversation with Dr. Robert Wachter below:
8 Ways to Navigate the AI Era in Healthcare (As Patient, Clinician, or Leader)
Stay in the Loop: Insist on human oversight for high-stakes decisions—diagnoses, treatments, bad-news delivery. Ask: “Is a human reviewing this AI output?”
Become a Skillful Prompter: For clinicians and empowered patients—learn to craft precise queries. Turn AI into a thoughtful consultant, not an oracle.
Demand Explainability: When AI suggests something, ask “Why?” Push for tools that show reasoning, not black-box confidence.
Guard Against Bias & Hallucinations: Cross-check outputs. Use AI as a second (or third) opinion, never the only one.
Advocate for Time-Restoring Tools: Prioritize AI scribes, inbox triage, prior auth automation—anything that gives clinicians back hours for human connection.
Embrace Patient Empowerment: Use AI for preliminary research, but verify with professionals. Become the CEO of your health—data in hand, questions ready.
Watch the Incentives: Support policies that reward better outcomes, not just volume or speed. Call out misaligned reimbursement.
Hold Onto What Remains Human: Presence. Empathy in uncertainty. Moral weight of decisions. These don’t automate—they evolve.
Download the FREE Companion Guide & Digital Workbook here.
Which part of healthcare’s AI moment excites or worries you most?
What small step could you take this week to stay human-centered amid the tech wave?
Share in the comments—I read every one, and your reflections always deepen mine.
P.S. If this resonates, share it with a clinician, patient, or policymaker who cares about healthcare’s future. And if you’re not subscribed, hit the button—it means the world.
Read A Giant Leap: Available wherever books are sold.