Texas has built a $25 million hub to supervise artificial intelligence across its public hospitals and medical schools. The University of Texas System calls the effort UT REAL Health AI.
Two sources fund it: $15 million the state legislature approved in 2025, and $10 million from the system’s Board of Regents. The project extends a Health Intelligence Platform that dates to 2016.
The footprint is large. Thirteen campuses take part, among them eight healthcare institutions and seven medical schools. Roughly 130,000 clinicians, faculty, and staff now fall under the new governance rules. MD Anderson Cancer Center and the UT Medical Branch are included.
Some tools already run in production. At UT Southwestern, Andrew Jamieson’s team adapted GPT-4 and Google Gemini to grade recorded patient exams. Humans keep the final say. Faculty make pass-fail calls, and disputes route to a physician. In a fall 2025 trial with 222 students, the model matched human raters most of the time and trimmed grading workloads by roughly 980 hours. A $300,000 grant will carry it to five more schools.
Two grants of $500,000 each back separate bets. UT San Antonio’s BEACON trains a triage assistant on hundreds of thousands of trauma cases drawn from seven hospitals. UT Austin’s WombWatch turns a phone into a fetal-movement detector using acoustic analysis.
Oversight is the harder half. Reporting by the Houston Chronicle flagged open questions about vendor terms, protection of patient data, and how AI mistakes get handled. Program leaders counter that few health systems anywhere are attempting the same level of transparency.
Across UT hospitals, staff have logged 571 AI uses. About 150 have moved into everyday operation.
