Newswise — Johns Hopkins nursing economist: The question isn’t whether AI should create efficiencies. It’s who captures their value.
As proposed billing and payment mechanisms for clinical artificial intelligence intensify debate across nursing, much of the conversation has centered on a striking comparison: Healthcare is developing ways to financially recognize clinical work performed by algorithms while much of registered nurses’ work remains difficult to identify within existing payment systems.
Olga Yakusheva, PhD, an economist and professor at the Johns Hopkins School of Nursing, argues that the comparison exposes a real problem—but risks leading nursing toward the wrong solution.
Rather than focusing primarily on preventing AI from being reimbursed or creating equivalent billing codes for nurses, Yakusheva says policymakers should be asking a different economic question: If AI creates savings or releases clinical capacity, who captures that value?
AI that safely automates administrative or routine work could create meaningful benefits for healthcare. It could reduce costs, release clinician time, and allow nurses to redirect their expertise toward patient education, care coordination, chronic disease management, transitions of care, and other unmet needs.
But none of those benefits is automatic.
A hospital could use AI-generated efficiency to expand patient care. It could increase nurses’ workloads or reduce staffing. Financial gains could be reinvested in the clinical workforce—or absorbed into operating margins and payments to technology vendors and investors.
Yakusheva argues that this allocation question deserves greater attention as policymakers determine how healthcare will pay for AI.
It also changes how healthcare should calculate AI’s return on investment.
Released nursing time, for example, is not necessarily a financial saving. If an AI tool reduces documentation time and nurses use those hours to provide more patient care, the technology has created additional clinical capacity. If the same efficiency is followed by reductions in nursing positions, it has been converted into a labor-cost saving.
Similarly, AI may eliminate one task while creating others, including monitoring algorithmic output, correcting errors, communicating results, coordinating follow-up, and managing safety concerns. Those downstream labor effects should be included when organizations calculate whether AI has actually reduced costs.
Yakusheva says the next phase of AI policy should therefore move beyond whether technology can be reimbursed and address how AI-generated economic value is measured and governed.
Among the questions she says policymakers and health systems should be considering:
- How should healthcare distinguish verified AI savings from projected savings or newly created clinical capacity?
- When AI releases nursing time, should that capacity be returned to patient care rather than automatically converted into workforce reductions?
- Should health systems be required to disclose how AI-related savings are calculated and where those gains go?
- What policies could encourage health systems to retain, retrain, and redeploy clinicians whose work is changed by automation?
- Should some verified AI-generated gains be reinvested in staffing, workforce development, and patient-care infrastructure?
- Could AI ultimately expand humanistic nursing care rather than diminish it?
Yakusheva also brings a different perspective to calls for nursing-specific billing codes. Healthcare already spends substantial resources on nursing; the deeper economic problem is how those resources are identified, allocated, and connected to the value nurses produce.
The policy challenge, she argues, is not to protect nursing from technological efficiency. It is to ensure that efficiency strengthens healthcare’s capacity to care for patients rather than simply becoming a mechanism for extracting labor costs from the system.
Yakusheva is available to comment on the economics of AI adoption in healthcare, nursing workforce investment, AI-generated savings and clinical capacity, healthcare payment reform, and how policymakers can structure AI incentives around reinvestment in patient care.