Hospitals are adopting AI faster than they can test it safely. Baptist Health South Florida built the governance structure most of the industry still doesn't have.
A new report shows most health systems are moving fast on AI deployment. Their testing infrastructure hasn't kept pace with that enthusiasm.
The Gap the Industry Is Facing
The report comes from UPMC's Center for Connected Medicine and KLAS Research. It surveyed health system leaders nationally about how they evaluate AI before rollout.
Less than half of hospitals have a dedicated environment for testing AI tools before those tools touch patient care. Validation methods vary widely, from formal vendor testing to loose pilot programs.
Sixty-three percent of health systems describe their own AI strategy as still developing or ad hoc.
Ken Howard, vice president of technology services engineering at UPMC Enterprises, points to a resource problem. Hospitals often lack the time or capital to build a structured testing environment first.
"They're going to go down that path just to learn that all that work potentially wasn't justified," he said.
What Baptist Health Built Instead
Baptist Health South Florida operates 11 hospitals across Miami-Dade, Broward and Palm Beach counties. It has already addressed part of that gap.
The system built ImpactAI, an internal portal listing approved AI tools, staff training and governance resources. It also ran a generative AI pilot specifically to reduce clinical documentation time for cardiologists.
"For healthcare providers contemplating the incorporation of generative AI technology, it's crucial to view AI as a tool that augments, rather than replaces, human input," said Douglas Davila-Pestana, Baptist Health's technical manager of AI.
That job title itself is a signal. Few health systems have a role dedicated specifically to AI governance.
A Research Partnership Adds Another Layer
Baptist Health's AI work extends into research, too. Dr. Tom Nguyen, chief of the Miami Cardiac & Vascular Institute, has partnered with FIU's biomedical engineering department on AI-driven cardiovascular research.
The collaboration uses AI to find patterns in patient data. The goal is diagnosing cardiovascular disease earlier and predicting post-operative complications through wearable sensors.
Why Ongoing Monitoring Still Matters Most
Building governance up front is only part of the equation. UPMC's own chief medical information officer, Rob Bart, stressed that oversight has to continue well past a tool's initial rollout.
Clinical algorithms predicting length of stay or readmission risk need regular rechecking. Bart said UPMC monitors these tools on set intervals to confirm their guidance still holds up.
Clinician use of AI tools nearly doubled between 2023 and 2026. That's according to a survey cited by DynaMed senior medical director Kate Eisenberg. That pace makes consistent evaluation standards harder to maintain industry-wide.
The Standard Other Hospitals Are Missing
Baptist Health's approach doesn't solve the industry's broader problem. Most hospitals still don't have a repeatable way to validate AI tools before or after deployment.
What South Florida's largest health system shows is that the infrastructure gap isn't inevitable. It's a resourcing decision, and Baptist Health made a different one.
Miami Luxury Retail Faces a New $9.5M AI Rival