America is facing a health care crisis that extends beyond rising costs. Physician burnout is increasing, access to care is deteriorating, especially in primary care settings and rural or underserved communities, and administrative burdens continue to squeeze time spent with patients. Doctors are expected to diagnose and treat illness in addition to navigating a complex system of documentation, payment rules and compliance requirements. This burden has become a major driver of staff reductions across the system.
To cope with the situation, healthcare organizations are turning to artificial intelligence tools as a practical necessity. AI scribes streamline clinical documentation, while AI-enabled coding solutions translate notes into accurate billing codes in real-time. This technology allows doctors to focus more on the delivery of care and less on paperwork, a result that almost all participants agree is late.
The performance benefits are clear. AI coding systems don’t just stop at reducing the administrative burden; they can also improve financial performance. Mercyhealth, for example, reported a Revenue growth of 5.1 percent after using the AI code solution. Health systems that use automated coding also see a significant reduction in claim denials, a problem that can cost large organizations. up to $5 million per year. At a time when hospitals are operating on razor thin margins, this efficiency is no small advantage.
But as providers increasingly use AI to stabilize their operations, payers are reacting with suspicion. Insurers have begun to identify the use of automated coding as “overcoding,” and executives at major companies, including UnitedHealthcare and Centene, recently signed plans to deploy additional AI tools to oppose what they describe as aggressive payment practices. The result is an emerging AI arms race across the revenue cycle that risks deepening mistrust rather than fixing the underlying problem.
It’s about the plot. The US health insurance model is built on utilization management practices that deny, delay or reduce claims as a means of controlling costs. Although insurance plays an important role in society, its economic benefits are fundamentally disproportionate to those of providers and patients. In response, doctors and health systems have been forced to write and code with extraordinary precision just to get paid for care already delivered. What could have been a straightforward process has turned into a system defined by complexity, lack of transparency and inconsistent rule changes.
In this situation, manual payment and coding is no longer a reality. The volume of documentation requirements, rule reviews and coding reviews exceeds what even highly trained coders can reasonably manage. AI is not a shortcut or an income manipulation tool. It’s the only scalable way to navigate a payment ecosystem that’s too complex for humans to see. In the modern healthcare environment, AI has become a basic infrastructure.
This tension is exacerbated by the slowness of regulation. Most US reimbursement systems—especially those developed by the Centers for Medicare & Medicaid Services—are designed for manual, person-coded time. Yet those same rules now govern AI-assisted workflows without updated guidance on how automation should be tested, evaluated or encouraged. Without modernization, the policy risks punishing efficiency rather than rewarding accuracy, leaving providers stuck between outdated compliance standards and operational realities.
As a practitioner with first-hand experience, the concern that AI-enabled code exists to raise debt misunderstands both the technology and the problem it aims to solve. Good coding is not about decoration. Automated systems ensure that services provided are captured correctly the first time, reducing the need for rework, complaints, and long return cycles.
A few years ago, I became a senior physician in a dementia care facility, to fill gaps in the care of residents who could no longer be seen by their physicians. Despite enjoying the experience of caring for patients in their skilled nursing facility, I found it difficult to code for care delivered outside of my office. After nine months of endless delays and denials, I realized again that I was earning about 25 cents on the dollar per patient visit. So, I “burned out” before a year passed and resigned, due to the coding burden.
Expecting providers to accept partial payment for care delivered would be like asking a grocer to allow shoppers to leave their store with only half of what’s in their cart paid for. Return delays and denied claims impose far greater costs on the system than accurate coding ever did—costs that ultimately trickle down to patients through reduced services, longer wait times and, in some cases, facility closures. This year only, 23 hospitals and emergency departments are closed. From a policy perspective, accelerating hospital closures, particularly in rural and underserved areas, raises questions that go beyond individual balance sheets. Reimbursement delays, denials and diversions increasingly shape which communities end up accessing care at all. This is not a sustainable model for anyone.
Importantly, modern AI coding platforms are highly readable systems where all payment decisions are traced back to specific clinical records. This transparency provides a clear rationale for claims, which paves the way for greater accountability for both providers and payers.
Notably absent from this debate is a balanced examination of how payers are using AI themselves. Insurers continue to use automated systems to flag, delay or deny claims at scale, often with less transparency than provider-side tools. Controlling one set of algorithms while leaving the other untested increases asymmetry and mistrust. A constructive way forward will focus on shared values rather than blame. Establishing clear guidelines for AI-assisted coding that define audit requirements, document tracking and acceptable usage across payment systems and providers will replace scaling with a common accountability framework.
Payers and providers ultimately share the same stated goals: delivering high-quality care, efficiency and maintaining financial performance. Treating AI coding tools as ammunition in an ongoing war undermines all three. When used correctly, this technology offers the opportunity to simplify an overengineered system, reduce friction and refocus resources on patient care.
Ending the AI arms race will require a paradigm shift. Progress depends on cooperation and realizing that individualism can be a shared tool for transparency, fairness and sustainability. Without that reset, the system risks continuing in a way that exhausts doctors, frustrates institutions and leaves patients with few places to turn.