How AI Will Change Your Health

In the 1950s, Homer Warner, PhD, MD ’49, was working at what is now Intermountain Health LDS Hospital when he envisioned the potential of computers to improve patient care. Considered one of the fathers of health-care informatics, he pioneered seismic shifts in health care: from digital patient record systems to a model that could diagnose patients with congenital heart disease faster than physicians. He also founded and served as the director of the world’s first degree-granting Department of Biomedical Informatics, which was at the University of Utah.

“We’re in a unique period of humanity with the Fourth Industrial Revolution, for which we planted the seeds 60 years ago in our department,” said Yves Lussier, professor and chair of biomedical informatics, who helped establish the semantics to teach computers the language of medicine. “We have Homer’s same can-do spirit, and we’re still international leaders in electronic health records, in clinical decision support systems, and in demonstrating that predictive analytics in medicine can save lives.”

Indeed, that reputation earned Salt Lake City the honor of hosting the International Conference on Artificial Intelligence in Medicine in July 2024—the first time it had been held in person in the United States since its inception in 1985. Five members of the Spencer Fox Eccles School of Medicine served in critical positions in planning and organizing the event, including Lussier and Professor Joseph Finkelstein, who pioneered informatics-empowered Telemedicine.

Here, we share just a handful of the faculty and alumni who are building upon Warner’s and the U’s legacies of using technology—in this case, AI—to improve health.

Mining algorithms for cancer treatments

Professor of Neurobiology Christopher Gregg spent his career working on genomics, watching big data transform the field. Then, in 2018, he was diagnosed with Stage IV breast cancer—a rare form of cancer in men with only a five-year survival rate of 20%. There was little research and no clinical trials for this cancer in men.

Gregg knew that all cancers eventually develop treatment resistance, so he began to optimize the timing, dose, and sequence of his treatments and integrate knowledge from diverse fields—ranging from animal ecology and species evolution to nutrition and cancer biology—to build an improved care plan to battle drug resistance. But Gregg needed to make better decisions based on better data. He had been developing computer vision and AI technologies to measure hundreds of micro-aspects of natural behavior in his lab and could detect important biological changes in the mice with this approach. He knew that all clinical decisions for patients begin with their doctor asking questions to understand the patient’s symptoms and behavior. But the methods are subjective. He imagined that if technology could be developed to objectively analyze symptoms and behavior from precise measurements in patients, it would be a gamechanger.

In 2019, Gregg co-founded Storyline Health Inc. with Jeff Barson, a leader in smartphone and AI software development. The app allows providers to develop and deliver symptom and behavior assessments for scalable precision and personalized care. So, in addition to understanding treatment plans, his personal genetics, and lab results, Gregg began collecting his behavioral data via video recordings answering simple questions: How was he feeling that day? What symptom changes did he notice? The technology then makes over 20,000 measurements of his behavior, speech, and vocal patterns, building a model of his health. He is now nearing year seven of living with cancer and makes treatment decisions with his oncologist based, in part, on his Storyline data.

“The technology then makes over 20,000 measurements of his behavior, speech, and vocal patterns, building a model of his health.”

The technology is also being used for groundbreaking applications. The National Institute for Jail Operations employs it to prevent suicide and violence in US jails, while the Huntsman Mental Health Institute studies its role in predicting treatment-resistant depression and managing ketamine and psychedelic-assisted therapies. The Moffitt Cancer Center integrates it into cancer clinical trials. Moreover, Storyline’s technology is now the leading behavioral AI solution in non-medical fields, including law.

As for Gregg, he continues to use it daily, transforming a terminal diagnosis into what he calls a manageable chronic disease.

Mapping maternal health

Each year, roughly 1 in 175 pregnancies in the United States ends in stillbirth. For the women whose pregnancies resulted in loss—and their families—those numbers translate into very real trauma. And for many expectant mothers, the fear of becoming part of that statistic can result in undue stress.

“How much better off would all of us be if we could just avoid all of that unnecessary suffering?” asked Nathan Blue, an assistant professor in obstetrics and gynecology, who researches whether a fetus is growing at the prescribed rate, which remains the leading indicator of preventable stillbirths.

Blue believes our current standards for assessing pregnancy risk are flawed, as they assume all fetuses should be the same size and often overlook other important factors, such as genetics and environ mental risk factors. To address this, Blue and colleagues have developed an AI tool that examines not only where a fetus falls on the size chart, but also the mother’s medication use and health history, and the sex and anatomy of the baby, to determine precisely how big—or small—each fetus should be.

For Blue, some of the most interesting findings are around the fetus’ sex, something doctors don’t take into account even though ultrasounds can predictably detect it. Historically, female fetuses are both smaller than male fetuses and healthier, making them less likely to experience complications even when their growth is suboptimal. Yet, using AI, Blue and team have found that female fetuses are at higher risk than their male counterparts if a patient has Type I or II diabetes, the opposite of the typical influence of fetal sex. “These unexpected patterns could prove really meaningful,” said Blue.

“Instead of saying simply your fetus is too small, our hope is to say your fetus is small not because your placenta isn’t delivering enough oxygen or nutrients,” Blue said. “It’s because it’s a female fetus, and based on your family history and genetic markers, your baby is supposed to be this small. You can then avoid all the costs of monitoring, avoid an unnecessary labor induction, and not stress and lose tons of sleep.”

Seeing the future

If you have a smart watch, you can get notifications about everything from a spike in your heart rate and irregular rhythms to reminders to stand up, close your move ring, and take a deep breath. They’re all part of a trend to gamify our health.

But instead of gamifying health, Khizer Khaderi, MPH ’98, MD ’02, muses, “What if we healthify gaming?”

“Half of the world’s population plays video games,” he said, during a presentation at the 2023 Global Healthspan Summit. “Why not allow them to opt in to donate their data just as easily as you can donate your blood?”

A neuro-ophthalmic surgeon, futurist, and founding director of the Stanford Human Perception Lab, Khaderi is working to do just that through the Human Perception Project. The concept is simple: allow players to opt in, so at the end of every session, in addition to getting stats on their gaming performance, they’ll get stats on their Vision Performance Index, which measures their visual, cognitive, and motor performance. On the other end, researchers around the world will have access to psychometric and biometric data that can help them better understand everything from childhood development and overall health span to brain and mental health.

Khaderi calls it assistive intelligence instead of artificial intelligence and hopes that it will lead to a new kind of tool—beyond large language models—that can better understand what humans actually need.

“It taught me that the coolest technology shouldn’t be disruptive to somebody’s experience. It should complement or be completely passive and behind the experience.”

He’s already in talks with game developers and is no stranger to merging eye care with technology. Khaderi has served as a consultant and/or advisor to multiple organizations across the fields of gaming, spatial computing, wearables, and AI, including Riot Games, the NBA, MLB, Apple, Google, Meta, the Global Esports Federation, the IOC, and WHO—to name a few. After finishing his residency at the University of Arizona, where he worked with baseball and basketball players, he went on to launch his first startup, a neuro-modulatory video game company. He worked with major league baseball teams to create a home run derby-style game that simulated playing on a field—everything from the sound of the crowd to the glare of stadium lighting.

“There was no joystick, nothing fancy, the camera wasn’t on, but we found that if you played for at least three minutes, it could improve your batting performance by up to 22%,” Khaderi said. “It taught me that the coolest technology shouldn’t be disruptive to somebody’s experience. It should complement or be completely passive and behind the experience.”

Advancing surgery and surgical training

Imagine a future harnessing AI where, with the press of a button, custom guides are digitally projected onto an arthroscopic surgery image, identifying where to resect bone, place guidewires, or insert implants; where more precise and uniform procedures result in better outcomes; and where intraoperative imaging and videos with auto-generated labels and graphic masking of anatomy and pathology are automatically shared with patients and physical therapists, allowing for better-informed rehabilitation plans.

That is the goal behind Kaliber, an AI-enhanced platform that advances arthroscopic surgery while streamlining communication between patients, surgeons, and administrators.

“We have, most definitely, only begun to explore the potential applications of AI for surgery,” said Richard Angelo, MD ’85, chief medical officer for the company. “We have good reason to expect that AI will facilitate more effective surgical procedures, lead to fewer adverse outcomes, result in improved safety, and improve patient satisfaction.”

In 2019, the Arthroscopy Association of North America asked Angelo to learn more about Kaliber’s efforts to optimize surgical care by harnessing AI.

“In researching AI, it was clear that it has the potential to touch every aspect of our surgical practice.”

“In researching AI, it was clear that it has the potential to touch every aspect of our surgical practice,” he said.

Among the most exciting applications for surgeons, according to Angelo, will be the following: the development of guidance systems that are able to template key components of a procedure based on best surgical practices and techniques, the synergy of AI with robotic systems to autonomously perform specific segments of a procedure, the creation of pre-operative plans using MRI and CT data that model patient-specific anatomy and are available to the surgeon in the OR, and the ability to facilitate technique development and operative-skill assessment for future surgeons.

“We have the confidence of centuries of medicinal plant use in humans to be able to pick from an enriched pool of nature’s chemistry, which increases our chances of finding something safe, relevant, and biologically meaningful.”

Accelerating drug discovery, naturally

Centuries before Bayer chemist Felix Hoffman synthesized Aspirin in 1897, ancient Sumerians, Egyptians, Greeks, and Romans used willow bark to reduce pain and fevers. They knew the power of the plant, but it took much longer for researchers to pinpoint the exact chemical compound responsible for the therapeutic effect and transform it into an easy-to-replicate medication.

Imagine being able to massively expedite that process, discovering new molecules from natural sources that have the potential to become blockbuster medicines. There are countless reasons to look to nature for the next generation of drugs: over a third of all FDA-approved medicines are derived from natural sources, and that’s from the less than 1% of nature’s molecules that have been explored so far. That’s the premise of Enveda Biosciences, which leverages AI to understand and search through the vast, unexplored domain of nature’s chemistry.

“Our platform allows us to answer two main questions about all the molecules that exist in nature’s chemistry and search—similar to a Google search—for the molecules that have the highest likelihood of therapeutic value,” said Hannah Gordon, PhD ’15, who serves as chief product officer for the company. “These are then rigorously tested, modified, and turned into medicines that a doctor can prescribe.”

Part of the challenge in the past has been how slow and arduous it is to discover and characterize even a single new molecule from a natural source, such as a medicinal plant. Using machine learning and mass spectrometry, Enveda can speed this characterization—ie, what a molecule does and what it looks like—of a novel molecule from one at a time to 10,000 at a time. So far, it’s promising. In only four years, Enveda has identified numerous potential drugs, including one for skin conditions like eczema and one for inflammatory bowel disease. Both are slated to go into clinical trials within the next six months, with even more promising drugs on their heels. By merging the ancient wisdom of medicinal plants with the cutting-edge power of AI, the company is rewriting the future of medicine, condensing what once took decades into mere months.

“We have the confidence of centuries of medicinal plant use in humans to be able to pick from an enriched pool of nature’s chemistry, which increases our chances of finding something safe, relevant, and biologically meaningful.”

“We have the confidence of centuries of medicinal plant use in humans to be able to pick from an enriched pool of nature’s chemistry, which increases our chances of finding something safe, relevant, and biologically meaningful,” Gordon said. “By speeding drug discovery, my hope is it will make a very big impact on many, many patients.”