Award Setup Complete — Partner Site Planning Begins
With award setup materials submitted and infrastructure partnerships forming, the UT-REAL consortium transitions from setup to active Phase 1 planning.
Funded by UT REAL Health AI · University of Texas System
Scaling AI-Based
OSCE Assessment
Across Six UT Medical Schools
The MAPLES AI grading platform deployed across UT System medical schools via the UT Health Intelligence Platform, funded by the UT-REAL Health AI initiative.
LIVE
UTSW Production
6
UT Partner Sites
7,000+
Encounters Graded
Participating Medical Schools
Partnering across the UT System to validate AI-enabled clinical assessment.
Latest Updates
View all posts »Milestones, publications, and announcements from the project.
UTSW Pilot Surpasses 7,000 AI-Graded Encounters
The lead site reaches a major milestone with over 7,000 standardized patient encounters graded using AI in production, establishing the foundation for multi-site expansion.
Project Funded by UT REAL Health AI Pilot Program
Our project receives $300,000 from the UT System UT REAL Health AI Pilot Program to scale AI-powered simulation assessment across six UT medical schools over 18 months.
Six UT Medical Schools
Collaborating across the UT System to validate AI-enabled clinical assessment.
View all participating sites →Our Mission
Empowering Educators, Improving Doctors
Our mission is to give medical educators better tools so they can give students better feedback — faster, more consistent, and at a scale that wasn't possible before.
Share Best Practices for AI-Ready Rubrics
Share strategies and best practices for designing AI-compatible OSCE rubrics, while allowing each school to customize assessment to their own educational philosophy and clinical needs.
Scale AI Grading Infrastructure
Deploy the MAPLES AI grading platform across partner sites via the UT Health Intelligence Platform (UT-HIP), enabling automated, governed, multi-site assessment of standardized patient encounters.
Validate AI vs. Human Concordance
Conduct rigorous validation studies comparing AI grading accuracy against expert human raters across diverse patient scenarios and institutions.
Multi-Site Data Collection via UT-HIP
Build the largest multi-institutional dataset of AI-graded OSCE encounters through the UT Health Intelligence Platform (UT-HIP) — a HITRUST-certified, Azure-native platform with existing MOUs across all UT health institutions.
Cross-Institutional Collaboration
Six UT medical schools sharing rubrics, best practices, and lessons learned — building a community of practice around AI-assisted clinical education.
Publication & Dissemination
Publish findings in high-impact venues (NEJM AI, JMIR AI) and present at national conferences to advance the field of AI in medical education.
UTSW Pilot Results
Production metrics from UTSW — the foundation we're scaling across the UT System.
0.830
AI-Human Agreement (kappa)
7,000+
Encounters Graded at UTSW
6
Partner Institutions
$300K
Grant Award
Phase Progress
From single-site proof-of-concept to UT System-wide deployment.
Phase 1: UTSW Proof-of-Concept (Complete)
AI grading system developed and deployed in production at UT Southwestern. 7,000+ encounters graded, 3,200+ students assessed. Published in NEJM AI and JMIR AI. AI agreement (kappa = 0.830) exceeds human inter-rater reliability (kappa = 0.732).
Phase 2: Award & Planning (Current)
Funded by the UT REAL Health AI Pilot Program, March 2026 ($300K, 18 months). Award setup complete. Infrastructure engagement with UT-HIP and UTSW Enterprise Data Services underway. IRB template protocol in review. Site inventory and governance planning in progress.
Phase 3: Multi-Site Deployment
Phased onboarding of partner sites beginning with Wave 1 institutions. Site-specific technical audits, IRB/DUA routing, and data ingestion pipeline deployment.
Phase 4: Validation & Dissemination
Cross-institutional validation studies, publication of multi-site results, open-source governance playbook, and framework for national adoption.
Where It All
Started
Born at the UT Southwestern Simulation Center, the MAPLES platform has graded over 7,000 clinical encounters in production — proving that AI can match and exceed human inter-rater reliability. Published in NEJM AI and JMIR AI, with multimodal assessment research on preprint. Now, through the UT-REAL initiative and the UT Health Intelligence Platform, we're scaling that capability across the UT System.