A multi-university team that includes University of Texas at Austin researchers has been awarded a grant from the National Institutes of Health to create AI-based interventions to help address the risk for suicide among Black youth.
The project is in response to what has become a nationwide mental health crisis, as young people are found to be suffering from sharply rising rates of anxiety and depression. Today, suicide is ranked as the second leading cause of death among 10 to 24-year-olds in the U.S. The rates are higher for Black children than other children.
The research team, which includes faculty from the Moody College of Communication, the School of Information and the McCombs School of Business, hopes to identify why that is. They will work with researchers from Cornell University and in collaboration with Prairie View A&M and Tuskegee University in Alabama, two historically Black universities.
Using data from National Violent Death Reporting System Restricted Access Database, the team will develop natural language processing models that can scan through thousands of death reports to identify risk factors for suicide, including demographic, environmental, economic and social factors, among others.
“We suspect there are a variety of issues that combine to create a powerful, if not deadly, recipe for the outcomes we are seeing,” said S. Craig Watkins, a professor in both Moody College’s School of Journalism and Media and Department of Radio-Television-Film, who is part of the team. “Our goal in this project is to develop more relevant artificial intelligence and machine learning algorithms to address these real world health problems in a way that’s responsible, ethical and equitable.”
The team will conduct the research alongside community stakeholders, such as health care professionals, hospitals and nonprofits.
Historically, AI models such as these are created by research teams and then shared with community groups. The goal is to change that outdated model to include stakeholders from the beginning, to help identify the kind of questions that should be asked and what sort of information would be helpful to design interventions. Another goal is to build trust in AI models and get buy-in from stakeholder groups who will rely on the technology.
“These tools can help us better understand the individuals we serve,” said Larry Wallace, co-founder of the Black Men’s Health Clinic who will join the team’s advisory board. “If we can find out if people are on medications, living in a particular environment with different access to food and different social determinants of health, we can create a profile to help people find the best providers.”
The NIH gave the grant through its new Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity Program, an effort to find ways to use AI to help address health disparities. The project is led by School of Information professor Ying Ding and also includes McCombs professor Yan Leng, along with Watkins, who works with the UT research initiative Good Systems that strives to help design and implement ethical AI systems.
