As the University of Tennessee at Chattanooga School of Nursing continues to expand its simulation toolkit, artificial intelligence is making a significant impact.
Earlier this summer, students in the Doctor of Nursing Practice (DNP) adult gerontology acute care program piloted an end-of-life training scenario unlike anything they had previously encountered in the University’s Metro Annex Safe Hospital.
The twist? Students weren’t just caring for a simulated patient in the mock clinical setting. They also had to navigate a phone conversation with the patient’s daughter, which was played not by a faculty member or actor, but by ChatGPT.
The AI phone caller was programmed with a detailed emotional profile: an adult daughter who was confused, anxious and upset that her father had been placed on life support despite having advanced directives.
The DNP acute care students had to guide her through the crisis, explain medical decisions in plain language and respond to unpredictable questions, all while managing the patient’s clinical care.
“It felt very realistic,” student Alexis Murray said after participating in the simulation. “Especially having the AI that we were talking to. We didn’t know what it was going to ask us or say, so it does give you that complete surprise effect.
“You don’t know what’s going to come next, which is realistic in the real world.”
Her experience was exactly what faculty hoped the simulation would evoke.
UTC Clinical Assistant Professor Chris Doneski, who helped design and run the scenario, said simulations like this are about building readiness and reducing the likelihood that a new nurse practitioner will face a real-world, emotionally charged conversation for the first time without preparation.
“These situations can be very complicated; there’s a lot of emotion, a lot of different feelings, family conflicts,” said Doneski, an adult gerontology acute care nurse practitioner (AGACNP) who joined the UTC faculty in 2024. “If you haven’t ever been exposed to that, just being thrown in the fire for a first time can be very uncomfortable and very unsuccessful.”
The AGACNP concentration prepares students to care for adults with complex, acute or critical health conditions, often in settings such as hospital emergency rooms and intensive care units.
Murray, who received a bachelor’s degree in nursing from UTC in 2019, knows that uncomfortable challenge firsthand. She spent several years as a traveling ICU nurse before returning to Chattanooga and enrolling in the DNP program in fall 2023.
“As an ICU nurse, I have been in ‘more than I wish’ scenarios that are the same way,” she said. “But the roles are different once you’re a nurse practitioner. As the provider, it’s our job to explain everything that’s happening, talk to the family, evaluate all the key points and then ultimately help lead them to what’s going to be best for the patient.
“It’s a lot more strenuous in the provider role than it is in the nursing role.”
DNP acute care student Alexis Murray, under the watchful eye of Dr. Chris Doneski, talks with a patient’s daughter during the AI simulation. Photo by Angela Foster.
Doneski said creating the end-of-life simulation scenario took hours, from developing the patient case to designing the AI’s emotional responses.
He utilized a storyboard approach in laying out the end-of-life simulation “like scenes in a film,” creating a character script that outlined the daughter’s emotional state. He taught ChatGPT to convincingly play the role, practicing conversations to train the model and fine-tuning the responses to introduce realistic unpredictability and real-world situations.
“For me, it’s been a learning experience from a new technology standpoint,” Doneski said. “I’m a millennial; I have been through the whole start of the internet, things getting more widely used and now AI. This is not something that I utilized growing up.”
He previously helped build a custom software program to bring HAL® S5301, a Gaumard-produced manikin described as the world’s most advanced interdisciplinary patient simulator, to life.
HAL played the role of the father in the latest simulation, but “he’s technically kind of dead in this scenario—so he isn’t being too fancy,” Doneski said with a laugh.
The new AI simulation scenario also served as a tool for assessment. Using rubrics fed into the system, ChatGPT evaluated student responses while faculty reviewed performance from both clinical and emotional standpoints.
“ChatGPT objectively grades their conversation from the rubric that we gave it,” Doneski said, “and then we came together as faculty and discussed how we felt they did and had a combined grade.”
Alexis Murray, left, Dr. Christi Denton and Dr. Chris Doneski in the Metro Annex Safe Hospital.
The idea to use AI as a role-playing voice actor came after Doneski and UTC Assistant Professor Christi Denton, the coordinator of the AGACNP concentration, attended a presentation at the Tennessee Simulation Alliance conference earlier this year. The keynote speaker demonstrated how conversational AI could be used to simulate a child in an oncology ward, prompting Doneski to consider applying AI in end-of-life care scenarios.
“Especially with our students really wanting more exposure to palliative care and these end-of-life scenarios,” he said, “we started doing some research and found that ChatGPT does offer some voice recognition, conversational AI, things of that nature.”
Denton, a five-time UTC graduate who joined the faculty in 2018, also works shifts as a nurse practitioner at Erlanger Baroness Hospital. She said she has witnessed firsthand how far simulation has evolved in recent years.
“It has been an unbelievable evolution in seven years,” Denton said. “From ‘we’re going to get together and do procedures’ to this—where we’re utilizing AI to actually have students mimic conversations—it’s truly incredible.”
She admitted being skeptical about simulation early on, calling her initial experiences “clunky” and lacking realism. That changed when Doneski joined the program.
“Having somebody come in and go, ‘This is what we can do with simulation,’ opened my eyes,” Denton said. “I couldn’t be more thankful for what he’s done here.”
She observed the recent simulations on a monitor outside the Safe Hospital room, watching the AI adapt in real-time.
“Every student that came in took a different approach to how they talked to the daughter and it adjusted to it immediately,” she said.
Using AI in these types of scenarios instead of a human actor created a more consistent learning experience across the board.
“If we bring in a standardized patient actor, there’s a lot to catch them up on,” Denton said, “and they’re going to get hit with things they don’t know how to answer.
“AI simulation makes every student feel like their interaction was both individualized and catered to what their statements were. This is AI responding to you. It’s giving back what you are giving to it, so it makes it a really clean simulation.”
Murray said the experience reinforced the value of simulation.
“I think it’s extremely important to be able to practice any of the scenarios that we’ve had,” she said. “If we do mess up, then it’s a dummy. We can just start over again and don’t have to worry about anything. But it also allows us as future providers to build knowledge and confidence in ourselves.”
Murray said she approached the simulation as if it were real, even though she knew the patient—and the end-of-life pronouncement—was not.
“I try to at least make all of the mannequins real in my mind,” she said. “It does feel real. In Tennessee, we have two-nurse verification and the pronouncement of death, so I have pronounced people before. It doesn’t get any easier. And it did feel real in this scenario, too.”
Doneski deemed the new scenario’s pilot run a success, with plans underway to expand AI-supported simulations into other parts of the curriculum. As a faculty member who also works regular shifts as a nurse practitioner at CHI Memorial Glenwood in Hixson, the constant stream of real-world cases keeps inspiring new ideas.
“There’s always something new,” he said. “Especially for those rarer cases or things that we don’t see very often … That’s the beauty of simulation.”
Learn more
DNP Nurse Practitioner Program
Call me HAL: UTC nursing students learning from state-of-the-art patient simulator
Photo gallery by Angela Foster
Alexis Murray discusses the AI simulation with Nursing Simulation Coordinator Michelle Rosano.
