Artificial intelligence is supposed to revolutionize health care. Instead, early evidence suggests it's making patients pay more for less.
PwC projected this week that medical costs will jump 9% in the employer insurance market next year and 8.5% in the individual market. The culprit is partly AI-powered documentation tools that providers are using to capture every clinical detail. The real problem: those details are being coded in ways that justify higher billing complexity, and insurers are passing the cost to workers and families.
The core issue isn't that patients are getting sicker or seeking more care. The costs are rising because of what PwC calls "changes in coded severity, case mix and paid amount per claim." In plain terms, AI is helping hospitals and clinics extract more revenue from the same services.
Paul Markovich, CEO of Blue Shield of California's parent company Ascendiun, acknowledged the dynamic plainly this week. "Within the health care system, the current incentives are to do more and get paid for more," he said. Companies deploying AI are asking one question: "How can I use this to further my self-interest?"
The economics tilted toward this outcome long ago. American medicine runs on fee-for-service payments, rewarding volume and complexity rather than patient health. AI is a tool that amplifies whatever system you put it in. When that system pays for more billable events, AI accelerates the trend.
Bob Kocher, a venture investor at Venrock, predicted the squeeze will tighten. "AI makes any system more efficient, and since our health system is already super efficient at driving fee-for-service units of care and coding, it is going to drive up both and make health care spending grow even faster," he said. Most AI investment today focuses on revenue cycle management and developing new drugs that will be expensive additions to the market.
There is a counterargument. Hossein Estiri, an associate professor at Harvard Medical School, contends that the current wave of billing-focused AI will eventually fade as health systems recognize that real value comes from improving outcomes. "The real value is to improve patient lives," he said, expecting AI to enable preventive care and precision medicine that reduces disease and cost.
The competitive picture is also uneven. Large for-profit hospital chains like HCA, Tenet and UHS can invest aggressively in AI faster than nonprofit hospitals, potentially widening the gap between well-resourced and struggling providers. At the insurance level, UBS analysis suggests administrative AI will eventually spread across the industry, competing away short-term gains but not necessarily benefiting patients.
The wild card is whether AI adoption forces the health care system to abandon fee-for-service payment altogether. Value-based care, which pays for outcomes instead of per-service billing, has circulated as an idea for years without gaining real traction. AI could finally make such a shift unavoidable, but only if policy and market pressure align to require it.
Author James Rodriguez: "Until health care's payment incentives change, every powerful new tool will be weaponized to bill more, not heal better."
Comments