The pace of AI innovation in healthcare is accelerating faster than the adoption of the infrastructure needed to support it.
AI and advanced analytics promise transformative improvements in clinical decision-making, population health management, and operational efficiency. From predictive models in chronic disease management to AI-assisted triage and documentation workflows, these technologies hold enormous potential. Yet the infrastructure required to deploy them safely and effectively has not kept pace.
AI applications rely on high-quality, standardized data, and without interoperability, their impact is limited. Inconsistent data definitions, siloed information, and fragmented workflows can compromise outcomes, slow adoption, and erode trust among clinicians and patients alike. While health systems and technology vendors rush to embed AI into everyday workflows, the foundational frameworks for data quality, transparency, and governance are still catching up.
Federal health IT initiatives are increasingly recognizing the critical role of standards in addressing these challenges. The CMS Interoperability Framework, the Trusted Exchange Framework and Common Agreement (TEFCA), and other emerging policy efforts emphasize not just connectivity, but also the need for common rules, consistent data structures, and strong governance to support safe, scalable AI adoption.
At the center of these initiatives is FHIR, the widely adopted industry standard for structuring and exchanging electronic health information. FHIR has become a critical component of federal interoperability initiatives, offering a modern, API-driven framework that supports secure, scalable exchange.
Today, most major U.S. health systems have implemented FHIR-based APIs in at least some capacity, and CMS has mandated their use for all Medicare and Medicaid payers by 2027 to support patient access and digital innovation.
Yet widespread adoption remains uneven, particularly among smaller practices, rural hospitals, and community health centers. Closing that gap is essential if AI is to benefit all patients, not only those in resource-rich institutions. FHIR provides a foundation for creating AI systems that can operate reliably across diverse clinical environments. By supporting standardized data structures, provenance, and traceability, FHIR helps enable key requirements for integrating AI safely into clinical workflows.
But standards alone are not enough. Successful implementation requires coordinated action across the healthcare ecosystem. Providers must adopt workflows that can consume and act on structured data. Payers must make information available in standardized formats. Technology developers must integrate open APIs. Policymakers must align incentives to encourage consistent adoption. Only when these components work in concert will AI move from pilot projects to broad, sustainable deployment.
Public-private collaboration is also essential. Multi-stakeholder initiatives, spanning providers, payers, developers, and federal agencies, continue to offer practical insight into how AI can be operationalized safely. These efforts help establish shared expectations for data quality, clarify how models should be evaluated and monitored, and surface strategies for mitigating bias before it reaches the point of care.
Yet even the strongest collaborations cannot overlook a core truth: equity must be deliberately built into the AI ecosystem. Without intentional design and the right infrastructure, AI risks reinforcing long-standing disparities. Large academic centers and integrated delivery networks often have the resources to adopt and govern these tools, while smaller practices, rural hospitals, and community health centers may struggle to keep pace.
Standards such as FHIR provide one path forward, offering a common, accessible foundation for data exchange and innovation that can help level the playing field. When data is structured consistently and shared reliably, AI becomes more portable, more explainable, and more adaptable across diverse care settings.
Over the next decade, AI will influence nearly every dimension of healthcare delivery, from how clinicians document care to how payers assess risk to how patients navigate the system. Ensuring that this transformation benefits all populations will require sustained investment in interoperable, widely adopted standards paired with robust, aligned policy frameworks.
If the healthcare system builds that foundation now, AI can move beyond promising pilots and isolated proofs of concept toward scalable, measurable improvements in care quality, safety, and patient experience. That is the opportunity in front of us, and the responsibility of every stakeholder shaping the digital future of healthcare
Charles Jaffe
Dr. Charles Jaffe is the Chief Executive Officer of HL7 International, the global data standards organization.
