AI in healthcare is advancing quickly. As Daniel Vreeman of HL7 recently wrote, health systems are racing to test large language models, predictive algorithms, and new digital workflows—but the structures needed to govern these tools haven’t kept pace. Governance and standards are no longer optional; they’re the foundation that will determine whether AI succeeds or fails in healthcare.
AI won’t scale safely without shared rules. Standards are often seen as background tech, but in reality, they’re public infrastructure. They provide the consistency and transparency needed for trust.
If every model is developed, monitored, and deployed differently, organizations end up with siloed systems, black-box outputs, and higher risk. That’s why standards for explainability, monitoring, and portability matter just as much as the models themselves.
Healthcare has done this before. The Trusted Exchange Framework and Common Agreement (TEFCA) showed it’s possible to align federal policy, technical standards, and industry participation around a common goal.
AI governance needs a similar model. Standards should cover not only how models are built, but also how they’re deployed and monitored. This involves defining metadata, provenance, and context for model outputs, enabling clinicians to use them with confidence.
AI carries the risk of widening existing gaps in healthcare. Large academic centers have the resources to manage AI safely. Smaller hospitals and clinics may not. Without governance that makes AI accessible across settings, we risk repeating the digital divide of the EHR era.
Equitable AI depends on more than good data. It depends on who can use the tools, under what conditions, and with what safeguards. Governance frameworks can level the playing field so that trust, transparency, and safety aren’t limited by budget or geography.
In our eBook, AI Governance, Compliance, and Risk Management for Healthcare, we outline practical steps to start building this foundation:
Map risks across clinical, operational, and financial workflows
Align AI programs with existing compliance structures (HIPAA, NIST, COSO ERM)
Build monitoring systems to detect drift and bias early
Involve diverse stakeholders to ensure equity from the start
Government can provide incentives and guardrails. But real progress depends on shared accountability between developers, providers, payers, and vendors. Many of the building blocks already exist. What’s needed now is the commitment to align and scale them.
AI in healthcare is moving fast. If we want it to be safe, equitable, and sustainable, the work of governance has to move just as quickly.
👉 Download our eBook [AI Governance, Compliance, and Risk Management for Healthcare] to see how your organization can start building this foundation.