The Trusted Exchange Framework and Common Agreement (TEFCA) has moved more than one billion health data exchanges, and the U.S. Department of Health and Human Services has committed $1.3 million to strengthen network oversight. The milestone lands as hospitals and health systems wrestle with how to govern artificial intelligence responsibly - a question that defined the agenda at the recent HIMSS AI in Healthcare Forum.
Governance gaps are the immediate obstacle
Panelists at the forum warned that AI tools, even well-designed ones, face serious deployment risk without mature data standards and early engagement from clinicians, IT teams, payers, and regulators. The consensus pointed to a hard reality: health systems that rushed AI pilots over the past two years are now hitting accountability walls, bias monitoring gaps, and uncertainty about what meaningful human oversight looks like as models embed deeper into clinical workflows.
The conversation reflects a broader reckoning. For many organizations, the central challenge is no longer whether to adopt AI but how to build the structural foundations that let it function safely. That need for governance frameworks is pushing health systems to think beyond individual use cases and into the operational routines that surround them.
Providers take a deliberate approach to use cases
HIMSS CEO Hal Wolf told Healthcare IT News that hospitals are sequencing their AI investments carefully, starting with clinical documentation and supply chain management. Those lower-risk domains let organizations build competency before moving AI closer to the point of care. A chief medical informatics officer at Cincinnati Children's told the publication that the bigger AI opportunity is not adding new applications - it is using automation to strip out unnecessary steps and reduce administrative burden on clinical staff.
That incremental mindset aligns with what operational leaders are saying. The pressure to show quick return on investment is steering many teams toward friction reduction rather than splashy diagnostic tools that require lengthy validation. This measured approach also dovetails with the work of building AI for Healthcare governance structures that can scale.
Data quality, not model complexity, is the limiting factor
Aquila Health CEO Dr. Jaime Bland framed the core obstacle plainly: "the limiting factor in healthcare AI is not the sophistication of the model, it is the quality and consistency of the underlying data." Health records remain fragmented across systems and governed by inconsistent standards, which restricts what any model can reliably produce - regardless of its architecture.
The TEFCA milestone gains weight in that context. Interoperability infrastructure that moves clean, standardized data between organizations is effectively a prerequisite for AI that works at scale. The HHS investment signals recognition that the exchange layer needs ongoing attention, not just an initial buildout. That federal commitment also highlights the growing overlap between public health infrastructure and AI for Government oversight responsibilities.
Workflow friction emerges as the next frontier
Clinical IT leaders are increasingly arguing that healthcare workers do not need more tools; they need fewer unnecessary steps. AI that quietly removes redundant documentation tasks, flags duplicate orders, or streamlines care coordination may deliver more measurable value than high-profile diagnostic models awaiting clinical validation. Mercy, one of the larger U.S. health systems, is applying product development principles to patient navigation to connect people to the right care more efficiently.
Internationally, the same themes recur. Spain's National Health System is developing a collaborative AI development approach as a potential model for the European Health Data Space, while Nordic systems are drawing attention for modular digital architecture strategies that manage transformation without disrupting clinical continuity.
Why this matters for healthcare professionals
The data infrastructure that underlies AI ambitions is getting more durable, but the next pressure point is whether governance frameworks can mature fast enough to keep pace with deployment. For clinical and IT leaders, that means every AI initiative in 2026 will need to answer not just what the model does, but who monitors it, how bias is detected, and where human accountability sits when the tool is embedded in a care workflow. The organizations that treat those questions as design requirements rather than afterthoughts will be the ones that actually realize AI's operational value.
Your membership also unlocks:
