AI’s most important contribution to healthcare is removing the administrative drag that prevents clinicians from doing the work only humans can do. That distinction is essential. In an aging society, the binding constraint is human presence, not data processing power. No algorithm can change that fundamental fact. The question is whether AI will be used to expand that presence or to shrink it. 

The unit of value in healthcare is an hour of genuine, human clinical attention, and the system is losing those hours to paperwork.

The Demographic Crisis 

American healthcare is facing a demographic supply and demand crisis. According to U.S. Census Bureau projections, by 2030, one in five Americans will be over 65, sharply increasing demand for clinical care. At the same time, the supply of people willing and able to deliver that care is not keeping pace. 

The strain is already visible. Physicians work an average of 57.8 hours per week, yet only 27.2 of those hours, less than half, are spent on direct patient care. Nearly half of clinicians’ time is consumed by documentation, prior authorization, order entry, and insurance administration.  

Hospitals spent $43 billion in 2025 simply trying to collect payments already owed for care already delivered. Meanwhile, 133,000 licensed physicians are not engaged in patient-facing roles at all, absorbed instead into administrative or non-clinical functions. 

Forward-thinking healthcare organizations are already harnessing AI to free up physician time for direct patient care.

The Question That Needs to Change

Corporate narratives often frame AI primarily as a mechanism for reducing labor costs. Robert Shiller’s narrative economics is instructive here: stories shape decisions, and the labor-as-cost story directs investment toward headcount elimination even when the deeper problem is a shortage of available clinical hours.

The healthcare organizations getting AI right are asking a different question: How much more can our existing clinicians accomplish if AI removes what prevents them from fully doing their jobs?

That shift in framing treats clinicians as the scarce, high-value resource they are — not as costs to be minimized. Some tasks can only be done by nurses. Some tasks can only be done by doctors. AI cannot help a patient stand safely, provide a therapeutic presence, or build trust with a family navigating a serious diagnosis. The unit of value in healthcare is an hour of genuine, human clinical attention, and the system is losing those hours to paperwork.

Two Languages, One Company

The tension between labor-enhancing and cost-cutting AI can play out within the same firm, often in the same earnings call. CVS Health Corp. is a good example, precisely because it does not fit neatly into either camp.

In its 2026 proxy statement filed with the SEC, CVS describes deploying AI into pharmacy dispensing workflows to interpret prescription directions, historically a labor-intensive manual task. Tony Ambrozie, Senior Vice President and Chief Digital and Technology Officer, states the goal plainly: “freeing up our clinicians’ time to better serve patients.” CVS reports that AI gives Aetna nurses back roughly 90 minutes each day, time that can be devoted to member care and closing critical care gaps. Separately, Aetna launched an AI-powered claims platform that has reduced claims processing time by more than 20%. Management cited their explicit goal of reducing administrative burden on providers, so they can focus on patient care. These are labor-enhancing deployments in the most direct sense: AI removes friction, so clinicians can do more of what only clinicians can do.

CVS also reports reducing call center volume by 20% to 30% through AI, and its investor materials — presented at its December 2025 Investor Day — describe mid-teens adjusted earnings growth through 2028 driven by “cost discipline, margin improvement, and technology-enabled productivity.” The company is investing $20 billion in digital innovation, and the financial returns, not just the clinical ones, are front and center in investor communications.

CVS’s narrative shows that AI can be both labor-enhancing and cost-saving simultaneously. The question Shiller’s framework forces us to ask is which story becomes the organizing logic for investment decisions. When a 90-minute time return to Aetna nurses appears alongside a 20–30% call center reduction in the same annual report, which number shapes the next budget cycle? Which metric gets presented to the board when the two objectives come into tension?

Japan’s Proof of Concept

Japan offers the clearest real-world test of whether technology can substitute human care labor. Facing demographic pressures earlier than the United States, Japan invested heavily in care robotics, including lifting robots, therapeutic companions, and humanoid assistants.

A landmark Stanford and NBER study found that robot adoption reallocated human care workers’ effort instead of reducing their need. Robots handled monitoring and mobility assistance while humans shifted toward relational care that machines could not perform. Human presence did not decrease; it just shifted toward the work hardest to automate: comfort, judgment, trust, and relationship.

Japan revealed the irreducible core of care work: Remove every task a machine can handle, and what remains is still human.

AI Can Free Unique Human Work

Cost savings and labor enhancement are not mutually exclusive. The best-positioned firms will walk that line deliberately, using AI to free clinicians for the judgment-intensive, relational work that humans do uniquely well, while letting AI absorb the administrative and transactional burden that consumes so much of a clinical workday.

The question is whether efficiency gains are being reinvested in clinical capacity or simply retained as margin. Technology strategy can also be aimed at augmenting the human care owed to every patient.

---