Beginning this month, and continuing throughout 2024, the IC² Institute will prioritize research and programming that explores the intersection between AI and health, with a particular focus on underserved populations and safety net providers — clinics, agencies and practitioners that deliver health-related services to the uninsured and other vulnerable communities.
The Institute’s 2024 focus on AI and health fits synergistically with the University of Texas Year of AI initiative. The University recently announced that 2024 would be “the Year of AI” — a year of prioritizing and highlighting AI research and education throughout the UT community.
“There’s so much groundbreaking work related to AI happening across the UT campus right now … We’re excited to play a small role in that, probing the potential benefits and risks of deploying AI in health care, an arena with profound implications for well-being,” said IC² Institute Executive Director S. Craig Watkins.
The use of AI in health holds great promise; its proponents cite its potential for delivering transformative improvements in predictive, prevention-oriented, and highly personalized (precision) health care. Despite its potential for good, AI systems could perpetuate existing inequities.
AI models trained on past data will reflect data biases — so if, for example, the data used to train algorithms is collected from a mostly white population, the resulting model may be biased against people of color. There’s also a concern that AI’s hyper-focus on medical data could take away from our growing knowledge about social determinants of health (SDoH) – though it’s possible that can SDoH’s can be successfully threaded into AI systems.
And then there’s the question of access: Will the clinics and community health organizations serving vulnerable populations have access to the latest developments in AI? Or will the lack of access further exacerbate disparities? The IC² Institute will be investigating these kinds of questions.
The “AI and Health” focus builds upon the Institute’s legacy of technology development as well as the Institute’s recent work in the health disparities space. Last year, partnering with a variety of local and statewide collaborators, IC² explored ways to deliver health care in more equitable ways — including creating a more diverse health care workforce; finding more precise ways to understand social determinants of health, and brainstorming novel approaches to delivering health care resources in nonclinical settings such as public libraries.
The Institute has several AI and Health initiatives underway. In partnership with the Episcopal Health Foundation, Institute researchers are conducting an in-depth statewide survey of health practitioners to better understand the risks and opportunities associated with developing and deploying AI to serve the underserved. Further, the Institute is partnering with the Dell Medical School to fund UT faculty research related to the design of Health AI to mitigate health disparities.
In April, the Institute will host a conference, “AI Health for All” — bringing together local and statewide clinicians, funders and researchers to tackle subjects such as how we might use AI to mitigate the country’s maternal health crisis.
An IC² team led by Watkins and Senior Research Scientist Matt Kammer-Kerwick is working with a large NIH-funded dataset to develop computational models that analyze the social determinants of health with more empirical precision.
UT students will have opportunities to contribute to the Institute’s AI in Health research through capstone projects conducted in collaboration with the Design in Health program and the College of Pharmacy.
The Institute’s AI and health focus is part of a broader goal of innovating health and well-being — a strategic direction set by Watkins in 2022, when he became executive director of the Institute. Under this broader initiative, IC² is probing various dimensions of health and well-being, including ongoing social and technical shifts in health services (for example, telehealth and remote patient monitoring) and the design of ethical tech solutions.