The healthcare industry is bracing for a battle over whether and how clinical artificial intelligence should be funded.
In late September, the Food and Drug Administration authorized 1,357 AI-enabled medical devices. But very few of these tools are actively paid by insurers.
Some health policy experts and clinicians I don’t see that as a problem.
“With AI, a lot of the conversation is about how we get paid for each individual technology,” said Ateev Mehrotra, chair of health services, policy and practice at the Brown University School of Public Health. “If I could wave a magic wand, I would change our paradigm to: ‘How can we use AI to improve clinician productivity and efficiency so they can care for more patients, with high quality, and at lower cost?’ »
Others, particularly in industry, worry that the lack of payment will slow adoption and prevent useful AI from reaching patients.
In 2026, the AI payment debate will intensify as more devices enter the field. So far, only three AI devices have received a permanent CPT code from the American Medical Association, a crucial step in obtaining payment from Medicare and, therefore, most private insurers. But there are many more waiting in the wings: WADA has assigned more than 20 temporary Category III codes to AI, many of which will eventually become permanent.
In anticipation of the coming flood, the AMA is considering a potential new code category for AI, including tools that require no intervention from a doctor. It and other medical companies are quickly trying to advocate for physician-friendly reimbursement models.
Meanwhile, the Centers for Medicare and Medicaid Services are grappling with their own payment paradigms for AI, which rely largely on vendors’ own assessment of their software. The senators also bill to formalize AI payment pathways.
These processes could take years to finalize. Meanwhile, “health systems are looking at AI and wondering how they’re going to pay for it,” said Pelu Tran, CEO of AI governance firm Ferrum Health. Here are some examples of how money will be distributed on a case-by-case basis across different clinical areas in 2026.
A fee-for-service signal: coronary plaque analysis
In January, Medicare will begin paying doctors a set a national rate of just over $1,000 for the use of AI that analyzes the type and amount of plaque in a patient’s coronary arteries. It is one of only three AI tools to have received a Category I CPT code from the AMA, and its reimbursement models will serve as an important data point as the organization and CMS continue to deliberate on payment strategies for AI.
Plaque’s previous technology, which is also paid more than $1,000 by Medicare, uses AI to calculate blood flow in the coronary arteries from a coronary angiogram. “We’re at a point right now where we’re seeing almost universal payment for the procedure,” Eric Rubin, principal CPT advisor for the American College of Radiology, said during a session on payment for AI at the Radiological Society of North America meeting in Chicago last month.
In comparison, “payment for plaque testing has been very uneven,” Rubin said. “It’s slowly becoming more uniform, but we’ll have to see how it progresses over time. »
This will depend in part on doctors’ understanding of when a patient might be a good candidate for the technology and how they document their condition so they can qualify for payment.
“A lot of doctors don’t know about prescribing CT and AI,” said Jacob Agris, vice president of product management at ConcertAI, which recently launched a product to facilitate the use and reimbursement of certain AIs, including coronary plaque analysis. “It can actually signal to them, ‘Hey, if you have these indications, you should consider this if it’s appropriate for your patient.'”
This type of workflow tool could help health systems obtain more AI-related claims reimbursements — a boon to their bottom lines, but potentially a burden on national health care spending. The first year of standardized coronary plaque receipt will help regulators and health systems determine where the right balance lies.
Charging patients: Breast imaging
Across the country, more women getting their annual mammogram now have the option to select an AI add-on to highlight suspicious lesions. But without reimbursement from insurers, it’s usually patients who pay, usually between $40 and $50.
“All of us doctors felt like moving to a self-pay model wasn’t our preferred approach,” said Greg Sorensen, chief scientific officer at major ambulatory imaging company RadNet, which charges patients $40 for its AI-based screening program. “We would have preferred that payers had accepted this from the start. »
Howard Berger, CEO of RadNet, said the AI-based screening program generates profits for the company. The company performs about 1.6 million mammograms a year, and about half of the women opt for the program, generating about $30 million in revenue, he said.
This year, the self-payment trend in AI for mammography will continue with a new class of algorithms. A small number of radiology centers currently offer AI add-ons that look for arterial calcification in the breast, for about $90. And Clairity Breast, an AI device that predicts a patient’s breast cancer risk over five years, will launch a pilot program at Beth Israel Deaconess Medical Center for $199, said Clairity founder Connie Lehman, who directs the Breast Imaging Research Center at Massachusetts General Hospital. One exception is an AI-based breast ultrasound interpretation, which has a temporary CPT code.
In 2026, rising expenses for women, who typically undergo mammograms annually, could put more pressure on insurers to cover certain AI applied to breast imaging.
“For reimbursement, we really need CPT codes for these AI products, because otherwise we’re going to authorize or enable patients who can afford it to get the AI, which has improved performance,” Sarah Friedewald, chief of the academic division of breast imaging at Mass General Brigham, said during a session at RSNA. “And that’s really not an equitable way to implement AI.”
Value-based experiments: opportunistic selection
In the absence of patient reimbursement or payment, health systems can still choose to invest in AI. They just need to believe that the technology will improve the efficiency or quality of healthcare enough to be worth it.
“If they feel, based on the published literature or their own internal experience, that this provides enough value for them, then I think we should use that as a sign, because they are paying out of their own pocket,” Mehrotra said. “If it doesn’t, it’s also a strong signal that this tool may not be adding enough value to our health system.” »
To this end, several health systems – typically large academic medical centers – are implementing opportunistic screening for health problems using existing radiology images. NYU Langone is experimenting using A CT scan to look for signs of osteoporosisand Emory Healthcare is developing an algorithm to search mammary arterial calcificationa signal of cardiovascular risk, in standard mammograms.
“I like to call opportunistic screening a rare triple win in the U.S. healthcare system, especially for AI,” said Hari Trivedi, co-director of the Health Innovation and Translational Informatics Lab at Emory University. Patients are informed earlier about their health risks, health systems can generate more revenue from patients referred for follow-up preventive care, and payers can save money in the long term by avoiding hospitalizations due to major health events.
Opportunistic screening programs are typically implemented in clinical research to determine whether detecting risky signs in images actually improves patient outcomes – the first hurdle in deciding whether a technology is worth using. But a side effect of these studies is that a health system should also be able to track their financial impact. Health systems may not report these results in the same way, but their results will still be worth tracking.