By Spencer Dorn

"Trust me, I'm a doctor." This refrain was typically enough to satisfy payers, administrators, and patients throughout the so-called "golden age of medicine" (1950-1980s), when physicians worked in small practices they owned, set their payment rates, and were highly regarded and rarely scrutinized.

Times have changed. Over the past several decades, various forces have eroded physician autonomy, leaving many dissatisfied and burned out. Artificial intelligence will likely further this trend. Let me explain how.

Increasing Accountability

For much of the 20th century, physicians practiced with few limits. They mostly ordered tests and treatments they felt were indicated and charged whatever they wanted.

Starting in the 1970s, public and commercial payors launched a series of models—most notably HMOs, PPOs, and ACOs—designed to reduce costs and boost quality. Each model has included peer-driven (not algorithmic) mechanisms that constrain physicians (e.g., prior authorization, medical necessity reviews, and various intentionally inconvenient processes) and increase quality reporting requirements.

Many of these activities are burdensome and restrictive. For example, physicians today complete an average of 43 prior authorizations each week. Likewise, primary care physicians report an average of 57 quality metrics for value-based contracts alone.

Rise Of Evidence-Based Medicine

In the 1970s and 1980s, researchers highlighted how doctors often made flawed decisions due to a lack of evidence, bias, and clinical reasoning errors. These realizations birthed the evidence-based medicine movement, pushing physicians to incorporate the best available research-based evidence into their practice.

While this is a noble goal, critics argue that EBM has contributed "to the creeping managerialism and politicization of clinical practice" and "cookbook medicine."

Digitization Of Healthcare

The shift from an analog to a digitized, interconnected world has reshaped how all Americans live and work, and physicians are no different.

The Health Insurance Portability and Accountability Act of 1996 (HIPAA) set data exchange standards and encouraged physicians to submit claims electronically to improve efficiency. Veteran healthcare analyst Jeff Goldsmith explained this "markedly accelerated the use of electronic data interchange in healthcare, to the great advantage of health insurers."

Similarly, in the 2000s and 2010s, health systems and medical practices implemented the electronic health records around which clinical medicine now revolves. They also adopted software products for various non-clinical physician activities, including credentialing, scheduling, compliance, policy training, and time tracking.

These software tools enable payors and administrators to monitor physicians' activities, ensure they comply with various requirements, nudge them to change their behaviors, bombard them with notifications, and even surveil their medical decisions.

Industrialization Of Medicine

Large provider organizations now dominate healthcare, with roughly 75% of U.S. physicians employed by hospitals, health systems, or corporate entities. Similarly, a few insurers have concentrated markets, allowing them to shape how physicians practice and get paid.

These organizations subscribe to varying degrees of Taylorism, a management philosophy built on the belief that there is one best way to do every task. Many physicians are now subject to what the anthropologist David Graeber called a "tyranny of managerialism," characterized by excessive administrative work and corporate management techniques.

Similarly, sociologist Allison Pugh explains in her book The Last Human Job that when large organizations overemphasize efficiency and control, they degrade the type of "connected labor"—work that requires an "emotional understanding" with another person—central to clinical medicine.

As an extreme example, many UnitedHealth-employed physicians recently accused their employer of forcing them to up-code encounters, see too many patients, and engage in practices that "squeeze profits out of patients."

How Artificial Intelligence Could Tighten The Screws

Technology journalist Brian Merchant explains that since the 1800s, management rather than workers have captured the gains from automation. He warns that bosses will use AI as leverage against workers, keeping pay down and demanding higher productivity.

Also drawing on history, the economist and Nobel laureate Daron Acemoglu similarly warns, "We face the risk of excessive automation: many workers will be displaced, and those who remain employed will be subjected to increasingly demeaning forms of surveillance and control."

Computer scientist and anthropologist Ali Alkhatib sees AI as an "ideological project to shift authority and autonomy away from individuals, towards centralized structures of power." Some non-healthcare corporations, like Amazon, already harness AI-based "bossware" to monitor their employees more closely.

Healthcare organizations could follow suit, automating what Jeff Goldsmith describes as pervasive mistrust of physician judgment. For example, health insurers may use AI to more easily over-scrutinize physicians' treatment decisions. A class action lawsuit alleged UnitedHealth Group was illegally using a flawed algorithm to deny rehab care to seriously ill patients.

Likewise, while many promise AI will liberate physicians from needless busy work, it could just as easily overload them with even more junk and nonsense.

Finding Balance

Self-Determination Theory holds that autonomy—the fundamental need to feel a sense of control and personal agency over our actions—is tightly linked to intrinsic motivation, engagement, and overall well-being.

Accountability measurement, evidence-based medicine, digitization, and industrialization have diminished autonomy to the point that many physicians now feel less satisfied, less motivated, and more like cogs in a wheel.

I do not believe physicians like me should have unfettered autonomy. We must accept a degree of oversight as part of serving the greater good. But ideally, it would be our patients—not a distant third party—holding us accountable for acting professionally, competently, and compassionately.

As technological historian Melvin Kranzberg famously said, "Technology is neither good nor bad; nor is it neutral." AI may significantly benefit physicians (and patients) by reducing drudgery, improving decision-making, and more. At the same time, AI could also further erode physician autonomy.

We all have agency. Let's work to ensure technology developers and healthcare organizations use AI to support rather than control physicians.