Article · AI governance · October 21, 2026
Most hospitals cannot produce a current list of the AI running inside their own walls.
Only 22 percent of hospital leaders are highly confident they could produce a complete AI audit trail within 30 days if asked. That finding comes from Black Book Research's survey of 182 hospital leaders, fielded in October and November 2025 and reported by Becker's Hospital Review. Confidence drops further at small hospitals, to 15 percent.
The audit trail is the symptom. The same survey found only 29 percent of hospitals have implemented and enforced policies covering AI model inventory, lineage, and sign-offs. Another 48 percent are still drafting them. Put plainly: most hospitals cannot say, today, which AI tools are live in their building, where the underlying model came from, or who approved it for what use.
Money tracks the gap. Black Book found hospitals plan to spend a median of 4.2 percent of their 2026 IT and quality or safety budgets on AI governance and safety. Large systems budget 6.8 percent. Small hospitals budget 2.3 percent. The hospitals with the least capacity to absorb an AI-related error are also spending the least to govern one.
Ask leaders what is actually in the way, and the answer is not disinterest. Forty-one percent point to limited vendor documentation: model cards and drift reports vendors either do not provide or do not keep current. Thirty-three percent point to unclear ownership. Nobody has decided whether IT, quality and safety, or compliance is the department that owns an AI tool once it goes live.
This is not a story about hospitals slow to adopt AI. A separate survey of 233 health systems, fielded with CFO interviews in the second quarter of 2025 by the Healthcare Financial Management Association and Eliciting Insights and reported by Fierce Healthcare, found 88 percent of health systems already use AI internally. Seventy-one percent have deployed it in finance, revenue cycle, or clinical functions. Only 18 percent of those same systems have a mature governance structure and a fully formed AI strategy, and three-quarters of that 18 percent report more than a billion dollars in net patient revenue. Mature governance is not evenly distributed. It concentrates where the balance sheet can support it.
The pattern across both surveys repeats. AI arrived faster than the paperwork meant to govern it, and the shortfall is widest exactly where the margin for error is thinnest: small and rural hospitals running on the leanest staff and budget.
Closing the gap does not require a new framework. It requires three documents most hospitals do not currently have on file: a current inventory of every AI tool live in the organization, a named owner for each one, and the vendor's documentation of how that tool's model was built and tested. None of the three requires a budget increase to start.
So here is the question worth asking this week: if your board called tomorrow and asked for a complete list of every AI tool running in your hospital right now, how long would it take you to produce one?
Follow along on LinkedIn.
Working through a decision like this one? Let's spend 30 minutes on it →
Leading a rural hospital or clinic? See what your state received, which rounds are open, and whether your project has a path. Rural Health Transformation funding by state →