AI Scribes Are Quietly Driving Up Healthcare Costs

Here's something insurers and hospitals rarely agree on: AI medical scribes are making healthcare more expensive. Behind closed doors, both sides admit it. In public, they can't agree on what to do about it.

A new STAT News investigation published today reveals a growing tension at the heart of healthcare AI adoption — the tools designed to reduce physician burnout are simultaneously inflating medical bills across the system.

What AI Scribes Actually Do

If you've visited a doctor recently, there's a good chance AI was listening. AI scribes are tools that sit in on patient-doctor conversations, automatically generating clinical notes, diagnostic codes, and billing documentation.

The pitch is compelling: doctors spend less time on paperwork and more time with patients. At UCSF (University of California, San Francisco), 70% of physicians now use AI scribes in their daily practice. Adoption rates at other major health systems are climbing fast.

Why this matters: When AI generates billing codes, it tends to capture every possible billable detail from a conversation — details that a rushed human doctor might skip. The result is systematically higher bills.

The Quiet Agreement

The Peterson Health Technology Institute brought insurers and providers together for a roundtable earlier this year. What happened behind closed doors was remarkable.

As Caroline Pearson from the Institute noted, when the cameras were off, the consensus was clear — AI scribes are increasing coding intensity across the board. Both sides acknowledged it privately, even as they publicly blame each other.

Health systems argue that AI scribes are simply capturing care that was always being delivered but never properly documented. In their view, insurers have been underpaying for years, and AI is finally correcting the record.

Insurers see it differently. They argue that AI-powered documentation tools are essentially optimizing for maximum reimbursement, generating more complex billing codes that inflate costs without corresponding increases in care quality.

The Numbers Tell a Story

The scale of AI scribe adoption is staggering. What started as a convenience tool for overwhelmed physicians has become a standard part of clinical workflow at major health systems nationwide.

The cost impact is becoming impossible to ignore. When AI consistently codes patient encounters at higher complexity levels, the cumulative effect across millions of visits translates into billions of additional healthcare spending. Insurers are watching their claims data and seeing the trend clearly.

Why this matters: Healthcare spending already accounts for nearly 20% of U.S. GDP. Even a small percentage increase in coding intensity across the entire system could mean tens of billions in additional costs passed on to employers, patients, and taxpayers.

Why Nobody Can Fix It

The fundamental problem is that both sides have valid points — and misaligned incentives.

Providers genuinely benefit from reduced burnout. Physicians who use AI scribes report spending less time on documentation and more time with patients. That's a real, measurable improvement in working conditions for an overworked profession.

But the same tools that reduce burnout also maximize billing. There's no easy way to separate the legitimate documentation improvement from the billing optimization. Is capturing a previously undocumented diagnosis better care or just better billing? Often, it's both.

Insurers could push back by tightening reimbursement rules, but that risks penalizing thorough documentation. Providers could voluntarily constrain their AI tools, but that would mean leaving money on the table. Regulators could step in, but the technology is evolving faster than policy.

What Comes Next

Several potential approaches are being discussed. Payers are exploring audit mechanisms specifically designed to evaluate AI-generated documentation. Some health systems are considering transparency requirements — flagging when notes were AI-generated versus human-written.

The most likely near-term outcome is a series of awkward negotiations between insurers and health systems, with both sides wielding data about what AI scribes are actually doing to billing patterns.

For patients, the implications are real. Higher coding intensity doesn't just affect insurer bottom lines — it flows through to premiums, copays, and out-of-pocket costs.

The Bottom Line

AI scribes represent one of the first cases where AI adoption at scale is creating measurable economic side effects that nobody planned for. The technology delivers on its promise of reducing physician workload, but it's simultaneously reshaping the economics of healthcare billing in ways that could cost the system billions. As AI tools become standard across more industries, this healthcare case study is a preview of the unintended consequences that come when optimization algorithms meet complex economic systems.