The push to establish billing codes for AI-enabled clinical analyses is intensifying a longstanding debate over how the healthcare system values nursing work.
The American Nurses Association is closely reviewing proposed federal payment policies for emerging technologies, but it cautions that several separate issues are being conflated: CPT code development, CMS reimbursement rules, and hospitals’ use of AI tools.
The proposals have raised a legitimate concern among nurses: Healthcare may be moving toward reimbursing technology while much of nurses’ work remains financially invisible.
“The issue is nuanced,” Sarah Gray, DNP, RN, senior vice president of nursing impact at the ANA, told Becker’s. “But the North Star is certainly clear that AI must be built with nurses, governed with nurses and used to support — and never replace — professional nursing judgment, while advancing patient outcomes and the value of nursing care.”
The American Medical Association has proposed a framework for “clinically meaningful algorithmic analyses,” which could create a pathway for certain AI-generated analyses to receive billing codes. The proposal has sparked concern across nursing because registered nurses generally cannot bill Medicare independently for their work.
Nursing services are commonly folded into hospitals’ broader inpatient payments rather than reimbursed as individual services. The prospect of creating a payment mechanism for AI has therefore sharpened questions about why the healthcare system has struggled to identify and reimburse nurses’ contributions separately.
The ANA is not calling for blanket opposition to the billing frameworks AMA or CMS are proposing.
Instead, ANA urges stakeholders to distinguish what the proposed codes would do now, what they could allow later and how future payment policies may affect nursing practice.
The codes are not autonomous — yet
The current proposals largely focus on tools that help qualified practitioners analyze clinical information, including diagnostic imaging. They do not represent a broad shift toward AI independently providing nursing care, according to the ANA.
That distinction matters because the phrase “autonomous AI” can suggest technology making clinical decisions without meaningful human involvement. The framework could accommodate more autonomous uses in the future, but the immediate proposals center on tools that assist clinicians.
The CPT coding process and CMS payment policy are also separate. CPT codes describe services and procedures, while payers determine who may bill for them and under what circumstances they will be reimbursed.
The ANA participates in the CPT process to ensure nursing perspectives are considered when proposed codes could affect practice. It is also reviewing AI-related provisions in CMS’ proposed 2027 Medicare Physician Fee Schedule and plans to submit comments before the September deadline, Dr. Gray said.
“We are already in this space,” she said. “This isn’t new.”
Dr. Gray also emphasized that virtual nursing, telehealth and AI should not be used interchangeably. Virtual nursing is generally a telehealth-enabled care model in which a nurse provides care remotely. AI may support that model, but virtual care is not inherently AI-provided or AI-driven care.
That distinction is particularly important as nurses’ concerns about technology receive greater attention. Recent labor actions have focused on AI’s potential effects on staffing, workload, safety and job security, expanding nurses’ resistance to hospital AI policies into a broader national debate.
“Virtual care is not the same as AI care or AI-supported care, and nurses’ roles will evolve with technology, and our profession has adapted and been agile in the past,” Dr. Gray said.
AI may reduce work — or simply move it
Hospital leaders often make the business case for AI around efficiency, time savings and lower administrative burden. For nurses, however, eliminating one task can create several others, according to Dr. Gray.
Nurses may need to redesign workflows, complete training, explain tools to patients, obtain consent, monitor outputs and respond when their clinical judgment conflicts with an algorithm’s recommendation. Health systems also need governance structures and processes for maintaining integrations between AI platforms and EHRs.
“The implementation of AI would ideally reduce the burden of care for nursing, not add to it,” Dr. Gray said. “But what we have seen in so much other technology that’s been implemented within healthcare is that it really is just shifting burden.”
The EHR offers a cautionary example.
Digital records became foundational to care delivery, but their adoption required substantial work from nurses serving as trainers, superusers, workflow designers and end users. That work was necessary, but it did not create an independent source of nursing reimbursement.
AI could follow a similar pattern.
Dr. Gray described AI as a process embedded in care rather than a single procedure performed on a patient. Hospitals do not bill for using an EHR itself. In the same way, AI may shape how a nurse documents, monitors or coordinates care without becoming a discrete nursing service that can be billed separately.
That creates a paradox: AI may generate reimbursable clinical output while the work nurses perform to implement, supervise and correct the technology remains hidden within broader hospital payments.
Ambient documentation is among the most common AI applications entering clinical settings, Dr. Gray said. Such tools may reduce manual note-taking, but clinicians must still verify their output.
Any time spent correcting errors, reconciling discrepancies or documenting why a clinician rejected an AI recommendation should be included when health systems calculate the technology’s return on investment.
“AI is not the be-all and end-all to solve all of the problems in healthcare,” Dr. Gray said.
Build with nurses, not around them
The ANA’s overarching message to hospital leaders is that nurses should be involved before an AI product is selected, not asked to validate it after major decisions have already been made.
That involvement should begin with identifying the clinical problem and continue through design, testing, procurement, implementation and optimization.
“Nurses aren’t against AI,” Dr. Gray said. “Nurses want to be involved in the development and the problem identification, so that we’re not just doing AI for AI’s sake.”
A technically sophisticated product can still perform poorly if it does not reflect how care is provided in a hospital unit. Nurses are positioned to identify where an alert may create duplication, where a consent process could interrupt care or where an algorithm lacks clinical context.
Bringing nurses in late forces hospitals to correct those problems after implementation.
“You can spend the money cleaning up, or you can spend the money on the front end and have the buy-in of nurses,” Dr. Gray said.
Some health systems are already handling AI implementation well because nurses have been involved from the beginning, she said. In those organizations, nursing participation is part of the technology strategy rather than late-stage validation.
Co-design may also ease concerns about job displacement. Dr. Gray acknowledged that some nurses fear increasingly sophisticated AI systems could threaten their roles, particularly as the technology advances quickly.
She does not expect AI to eliminate the need for nurses, but she does expect roles to change. Health systems will need more nurses in informatics, AI governance and workflow design. Nurses may also move into positions focused on evaluating outputs, identifying safety risks and integrating technology into patient care.
Nursing has repeatedly adapted to changes in medicine and care delivery. AI presents another moment in which the profession will need to evolve while protecting human judgment, advocacy and connection at the center of nursing.
“We do need more efficiency in healthcare,” Dr. Gray said. “But we also don’t want to take that human touch away from the profession.”
For the ANA, the billing-code debate is about more than whether a particular technology receives a CPT code. It is an opportunity to reconsider how healthcare recognizes the value nurses create, including the work required to make AI safe, effective and usable.
“We have an opportunity to do things differently,” Dr. Gray said. “But we have to work together.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.