Artificial Intelligence, Real Solutions
Posted on April 28, 2025
Great ideas are sometimes born from simple conversations.
In April 2024, Dr. Bret Webb, a professor of civil engineering, bumped into Dr. Bhushan Lohar, an assistant professor of systems engineering, in a Shelby Hall elevator and suggested that he attend a research mixer that was coming up.
Lohar did and was immediately greeted by Dr. Candice Holliday, an assistant professor of obstetrics and gynecology in the Whiddon College of Medicine, who organized the mixer on behalf of the Faculty Senate in collaboration with the Office of Research and Economic Development.
鈥淔aculty at the mixer were asked to put research interests on their name badges,鈥� Holliday said. 鈥淚 noticed that Dr. Lohar鈥檚 badge said artificial intelligence and was so excited because my sister wanted to connect with someone doing research in AI鈥�
She introduced Lohar to her sister, Dr. Nicolette Holliday, an associate professor of obstetrics and gynecology, and the two discussed the idea of using AI to help develop medical students鈥� communication and professionalism skills.
鈥淎t conferences that my sister and I have attended over the past few years, we have been hearing more about utilizing AI for medical education,鈥� Nicolette Holliday said. 鈥淚 was interested in using it to provide educational experiences to better prepare our medical students for the clinical environment in residency and beyond.鈥�
Their conversation led to defining a vague requirement of how an AI could provide more real-life conversations that doctors have with their patients, but don鈥檛 while in medical school.
鈥淚t can be difficult to include students due to the sensitivity of some topics OB/GYNs discuss with their patients,鈥� Nicolette Holliday said. 鈥淭he idea was that if we could use AI to provide simulated experiences on potentially difficult conversations on restricted health topics, it would benefit everyone.鈥�
Lohar began brainstorming and tech scouting for similar ideas but discovered that there weren鈥檛 any robust solutions available. Current training systems for medical education provide standardized scenarios with pre-programmed responses.
Lohar believed that AI could improve those systems. He approached South鈥檚 Office of Commercialization and Industry Collaboration and found help with the National Science Foundation鈥檚 Innovation Corps, which helped identify market potential and customers.
鈥淒ue to the recent advances in artificial intelligence, the field is ripe for significant advances and technological improvements,鈥� said Dr. Christopher Koczor, associate director of the Office of Commercialization and Industry Collaboration.
Lohar鈥檚 Advanced Systems Research Laboratory, consisting of himself and graduate research assistants Sudhanshu Tarale and Ninad Pandit, went to work.
First, they conducted about 15 interviews with several medical students and professors to understand what exactly they needed in the application and how to begin thinking of a potential solution using resources available to them in the lab.
鈥淥nce most of the requirements were captured, we started building the system architecture using system modeling language and model-based systems pattern language frameworks,鈥� Lohar said.
By mid-December, his team was able to provide a demonstration of AI agent technology to members of the OB/GYN department and another demo for Nursing Simulation in January of this year.
鈥淲e are continuing to upload funds of knowledge to the AI agent so that the avatars can provide appropriate responses and cues to the learner,鈥� Candice Holliday said.
Lohar鈥檚 team is now developing an upgraded version of the technology they plan to test this summer.
The Office of Commercialization and Industry Collaboration has filed a provisional patent application. Koczor sees a broad horizon for this technology.
鈥淭he training programs are not limited to medical school education. They can be used for training in just about any field where a trainee must combine lots of information into a single, actionable response,鈥� he said.
Lohar anticipates a license to commercialize this technology by the end of this year. In the meantime, they will be running trials with different scenarios and incorporating feedback into the platform.
鈥淲e鈥檙e just scratching the surface with this technology,鈥� Lohar said. 鈥淚t has numerous applications, from healthcare to human resources. It can help train first responders and law enforcement officers, members of the aviation industry and much more.鈥�
