Arizona State University launched Meletai in the fall 2026 semester at its new John Shufeldt School of Medicine and Medical Engineering, where students earn an MD and a master's degree in medical engineering. Meletai is a platform in which students practice clinical encounters with AI-generated video avatars that talk and behave like patients. It was built by ASU Enterprise Technology's AI Accelerator team with Dr. Jordan Coulston, assistant dean of clinical education. Earlier in 2026 ASU leaders described the same work, including the Case Creator workflow, under the name Praxis AI.
The problem. Medical schools train students with standardized patients played by actors. Actors are expensive to scale, cannot be repeated on demand, and limit how thoroughly each student can be assessed. Writing a realistic case with a believable persona can take hundreds of faculty hours, and cases must be updated as medical standards change.
Case Creator. Patient cases are produced by an agentic workflow on ASU's CreateAI platform in which several agents work together to build a patient's medical background, current condition and previous clinical appointments. In the February 2026 description by ASU CIO Lev Gonick, Coulston and Roger Kohler, director of AI Acceleration, a case template agent selects the clinical framework, visit-generation agents build a medically coherent sequence of encounters, a narrative unification agent keeps persona and story consistent, and a verification agent checks clinical accuracy and alignment with the grading rubric. Faculty and instructional designers review and refine each case before release.
The encounter. Meletai works like a telehealth visit. The student logs in, reads a short explanation of why the patient has come, and interviews an avatar generated with multimodal video and audio models. The simulation can supply vital signs, laboratory results, imaging and other clinical data on request, so the student must decide what to ask for, interpret the findings, explain them to the patient and choose next steps.
Assessment. Each encounter produces a full transcript that can be run through several layers of assessment at once, from basic communication skills for a first-year student to clinical reasoning for a third-year student. The AI evaluation scores an interaction across 25 areas, including how the student gathered and used information, reached a diagnosis, communicated uncertainty and involved the patient in the plan.
Status. ASU says Meletai is in use with students this semester. It has not published usage numbers, the specific models behind the avatars, or validation of the AI scores against faculty ratings.