Date: July 14, 2026
Series: The Business of Health with Chip Kahn
Host: Charles N. Kahn III
Guest: Caroline Pearson, Executive Director, Peterson Health Technology Institute (PHTI)
Introduction: The Hidden Conflict in the Back Office
In the high-stakes world of modern American healthcare, a quiet but transformative conflict is unfolding. While public discourse surrounding Artificial Intelligence (AI) in medicine often focuses on breakthroughs in diagnostics, drug discovery, or robotic surgery, the most significant immediate impact is occurring behind the scenes: in the administrative back office.
In the latest episode of The Business of Health, host Chip Kahn sits down with Caroline Pearson, Executive Director of the Peterson Health Technology Institute (PHTI), to dissect the "administrative arms race." As providers deploy AI tools to streamline coding, billing, and revenue capture, insurers are simultaneously equipping themselves with sophisticated algorithmic defenses to review, deny, and audit those very claims. This technological tug-of-war raises a critical, multibillion-dollar question: Does this rapid integration of AI actually lower the cost of care for the American patient, or does it merely accelerate the friction between those who provide care and those who pay for it?
Main Facts: The Anatomy of the Arms Race
The integration of AI into healthcare administration is no longer a futuristic concept; it is an operational reality. The core of the issue lies in the incentive structure of the U.S. health system.
- Provider Strategy: Healthcare systems are increasingly leveraging AI to automate the revenue cycle. This includes using Large Language Models (LLMs) to generate clinical notes, suggest diagnostic codes, and ensure that documentation meets the stringent requirements for reimbursement. The goal is to reduce "administrative burden," a massive driver of burnout among clinicians.
- Payer Response: In reaction to the perceived increase in "optimized" or "upcoded" claims, health insurers are deploying their own AI layers. These systems are designed to detect patterns, flag anomalies, and automate the prior authorization and denial processes.
- The Net Result: As both sides become more efficient at "playing the game," the question remains whether the cost of these technologies and the resulting bureaucratic speed-up actually provide value to the system, or if they simply consume capital that should be directed toward patient outcomes.
Chronology of the AI Integration
- 2023–2024 (The Pilot Phase): Early adopters in both provider systems and insurance companies began experimenting with machine learning for basic administrative tasks, such as automating patient scheduling and routine billing queries.
- 2025 (The Proliferation): Generative AI became widely accessible. Providers began using AI to automate the creation of clinical documentation to combat the time-sink of Electronic Health Records (EHRs).
- Early 2026 (The Response): Insurers shifted their AI strategies from simple rule-based denials to advanced predictive modeling capable of auditing millions of claims in real-time, leading to a noticeable surge in complex claims reviews.
- July 2026 (Current State): The "arms race" has reached a critical point of maturity. Independent evaluators, such as the PHTI, are now beginning to measure whether these tools yield clinical improvements or if they are simply administrative "activity" masquerading as efficiency.
Supporting Data: Beyond the Hype
Caroline Pearson brings a unique perspective to this debate. Through the PHTI, she and her team conduct rigorous, independent evaluations of digital health technology. Their work is designed to move beyond the marketing collateral of tech vendors and into the cold reality of clinical and economic data.
The PHTI Evidence Base
A landmark study by PHTI—referenced in the discussion—examined digital diabetes management tools. The findings were sobering: despite the technological promise, these tools failed to meaningfully lower the total cost of care. This finding serves as a cautionary tale for the current AI boom.
According to Pearson, the "administrative AI" tools currently entering the market are subject to the same scrutiny. If an AI tool reduces the time a physician spends on a chart by five minutes, but that time is immediately reclaimed by more aggressive administrative scrutiny from a payer’s AI, the net systemic savings are zero. Furthermore, the cost of implementing these dual-sided AI infrastructures adds a new layer of "tech debt" to the healthcare system that must be recouped through higher premiums or lower provider margins.
Official Perspectives and Expert Analysis
The Host: Chip Kahn
Chip Kahn, a veteran of healthcare policy and a senior visiting fellow at KFF, frames the discussion around the fundamental tension of the U.S. system. His research, spanning the American Enterprise Institute and the USC Schaeffer Center, consistently emphasizes that policy dictates outcomes. If the system rewards "activity" rather than "value," AI will simply be used to optimize for activity.
The Guest: Caroline Pearson
Pearson, whose background includes senior leadership at NORC and Avalere Health, advocates for a shift in how we measure success. She argues that the healthcare industry must demand evidence that AI tools improve the quality of care.
"We are seeing a rush to deploy tools that solve the immediate pain point of the individual user—the doctor who wants to finish their notes faster, or the payer who wants to catch billing errors," Pearson notes. "But we are missing the ‘so what?’ of it all. Does this make the patient healthier? Does this reduce the massive complexity that plagues our system, or does it make the complexity faster and more automated?"
Implications: The Future of Health Policy
The implications of this arms race are profound for stakeholders across the spectrum:
1. For Patients
The immediate impact on patients is a paradox. On one hand, administrative AI could lead to faster claim processing and fewer manual errors. On the other hand, if AI-driven denials become more prevalent and harder to appeal, patients may face increased hurdles to receiving covered care. The "black box" nature of some AI decision-making tools poses a transparency risk that regulators have yet to fully address.
2. For Providers
For hospitals and independent practices, the challenge is an escalating capital expenditure. Keeping up with the "insurance AI" requires significant investment in proprietary or third-party technology. Smaller practices, in particular, may find themselves at a disadvantage, unable to afford the AI tools necessary to fight back against payer denials, potentially accelerating the trend of physician practice consolidation.
3. For Insurers
While insurers may see short-term gains in controlling costs, they risk significant reputational damage if their AI systems are perceived as unfair or overly obstructionist. Furthermore, if insurers spend too much on their own administrative AI, they may struggle to keep premiums competitive, inviting increased government scrutiny and calls for stricter oversight of AI in health coverage.
4. The Policy Path Forward
Pearson suggests that the solution is not to halt AI development but to align incentives. If payment models moved more aggressively toward value-based care—where providers are paid for health outcomes rather than the volume of services—the incentive to use AI to "game" the billing system would evaporate. Instead, the incentive would shift toward using AI to identify high-risk patients early and manage chronic conditions more effectively.
Conclusion: A Turning Point
The conversation between Chip Kahn and Caroline Pearson serves as a vital reminder that technology is never neutral. It reflects the incentives of the environment in which it is deployed. As we stand in mid-2026, the healthcare industry is at a crossroads. We can continue to fuel an arms race that consumes more resources and deepens the divide between payers and providers, or we can use the power of AI to fundamentally rethink how we deliver, code, and pay for care.
The Peterson Health Technology Institute’s commitment to rigorous, independent evaluation will be a critical bellwether in the coming years. By peeling back the layers of marketing and examining the true economic and clinical impact of these tools, stakeholders can begin to distinguish between technology that drives systemic progress and technology that merely accelerates the status quo.
As Pearson concludes, the goal of a high-performing health system is to deliver better care at a lower cost. If AI cannot move the needle on those two metrics, it is not a solution; it is merely a more sophisticated form of the same old problems.
About the Series
The Business of Health with Chip Kahn is a weekly podcast dedicated to the complex, often opaque world of healthcare finance and policy. By bringing together the industry’s most influential voices, the series aims to connect the dots between corporate strategy, government regulation, and the lived experience of the American patient.
For more information on the Peterson-KFF Health System Tracker and ongoing coverage of AI in healthcare, visit KFF.org.
About the Author
