The Trump administration has moved from framing artificial intelligence as a policy priority to deploying AI in federal health programs.
In its National Policy Framework for Artificial Intelligence, the White House emphasizes priorities tied to U.S. competitiveness, regulatory consistency, and broader adoption across sectors. These include workforce development, national security safeguards, intellectual property protections, and a federal role in setting standards.
Within HHS, the focus has been on establishing the governance structures needed to support that expansion. The department’s December 2025 AI strategy emphasizes risk management, internal capacity building, and systems to track and evaluate AI use cases, including maintaining an inventory of applications and aligning implementation with broader federal requirements.
Access our slide deck outlining the federal AI policy landscape here.
At CMS, these policies are beginning to take shape in practice. The agency is positioning AI as a foundational capability across administration, oversight, and aspects of care delivery. In April 2026, CMS launched the first wave of its Health Technology Ecosystem tools, with more than 700 organizations pledging support and tools from over 50 companies available or in development. As noted in our story on the Medicare App Library, the initiative focuses on digital intake, data sharing, and patient-facing applications, reflecting how AI-enabled infrastructure is beginning to take shape at the program level.
CMS is also expanding the use of AI in the detection of fraud, waste, and abuse, as the agency shifts from a reactive “pay and chase” model to a proactive approach. AI tools are being used to flag suspicious claims in real time. Earlier this year, CMS Deputy Administrator and Chief Operating Officer Kim Brandt told NextGov that the agency’s Fraud Detection Operations Center has used AI to help prevent more than $2 billion in improper payments. Targeted interventions have also contributed to substantial reductions in problematic billing patterns, including a sharp decline in suspicious skin substitute claims following payment changes. The Wasteful and Inappropriate Service Reduction (WISeR) model, which Meghan Basler summarized upon release, will apply AI and machine learning to utilization review for services vulnerable to inappropriate use, while maintaining clinician review of non-affirmed requests.
Over time, these capabilities may extend further into payment models and care redesign. As Applied Policy’s Annie Tuttle has reported, the Accountable Care for Complex Conditions and Episodes (ACCESS) model tests outcome-aligned payment for technology-supported care in Original Medicare, with a focus on chronic conditions such as high blood pressure, diabetes, and chronic musculoskeletal pain. The model is designed to reward improved outcomes and efficiency, rather than the adoption of new tools alone, a distinction that may shape how AI integration is evaluated in future payment models.