Registry Maps ‘Fragmented’ Health AI Policy Landscape

Researchers at the Icahn School of Medicine at Mount Sinai have developed a registry to help health system executives navigate what they call a fragmented health AI governance environment.

Publishing their work in an npj Digital Medicine Perspective article, the authors noted that several tools already exist to help track AI-related law and policy, but most focus on a single jurisdiction or instrument type and are not designed for health-sector decision-makers, who lack a comprehensive and health-specific view of policies that shape AI design, deployment, and oversight.

The Health & AI Policy Index (HAPI) was developed to help meet that need by aggregating U.S. state and federal measures, sector-specific regulations, international frameworks, and voluntary standards into a single structured dataset. Entries are screened for AI and health relevance, tagged for key themes, stakeholders, and impact, and linked to source text, prioritizing official sources where available; users can sort and filter for these tags or use a Trends view that visualizes patterns over time.

The Mount Sinai researchers analyzed 240 healthcare AI-related policies published between 2016 and 2025 and found that governance efforts are developing through a patchwork of regulations, institutional guidance, technical standards, and policy initiatives rather than through a centralized system. The authors say this fragmented environment may create operational and compliance challenges for health systems attempting to responsibly integrate AI technologies.

To conduct the study, researchers used the Health & AI Policy Index to catalog and analyze healthcare AI-related policies published over nearly a decade. The framework was designed to help track emerging policy trends and better organize the rapidly growing body of AI policy activity affecting healthcare delivery.

“Health systems are increasingly recognizing that successful AI adoption requires more than just implementing new tools,” explained senior author Girish Nadkarni, M.D., M.P.H., chief AI officer of the Mount Sinai Health System, in a statement. “It also depends on strong oversight, internal governance structures, and clear accountability around how these technologies are used,” added Nadkarni, who also is chair of the Windreich Department of Artificial Intelligence and Human Health and director of the Hasso Plattner Institute for Digital Health, as well as the Irene and Dr. Arthur M. Fishberg Professor of Medicine at the Icahn School of Medicine at Mount Sinai.

The authors say that by organizing policies into a health-focused registry with consistent metadata, thematic and stakeholder tags, and simple impact classifications, HAPI could help health systems, developers, and policymakers see how diverse measures fit together, identify which instruments are likely to matter most, and recognize gaps that may warrant action.

“Questions around transparency, patient safety, and accountability are becoming central to the future of healthcare AI,” added Nadkarni. “Our work helps identify where policy efforts are growing, where gaps remain, and where additional coordination may be needed.”