Ole Miss, University of Memphis partner to address isolation in older generation
Social isolation and loneliness produce serious impacts on mental and physical health, and older adults are likely to experience these conditions. A new project aims to use artificial intelligence to help identify loneliness of older adults as they tell the stories of their lives.
Researchers at the University of Mississippi and University of Memphis have partnered to develop an artificial intelligence system that can listen as a person tells their life story and produce an autobiography that can be shared.
While listening to the stories, the AI will analyze vocal patterns and other indicators of loneliness and depression.
“Older people are highly likely to be disconnected from significant persons, including family members, neighbors and communities,” said Seok Won Jin, associate professor of social work at the University of Memphis and the project’s leader. “This can lead to huge psychological distresses including depression, loneliness and many other things, which is a huge problem, not only for the individuals, but also their community.
“We need to take care of them; that is our community’s responsibility.”
Around 1 in 6 people worldwide experience prolonged social isolation. The number increases to nearly 1 in 4 for adults over age 65 in the United States.
While earning his degree in gerontology, the study of aging, Jin once spent time in an assisted living facility, where he interviewed older adults and wrote biographies of their lives. The biographies were meaningful both to the individuals and to their families.
“But if we wanted to do this at a larger scale, we could not ask Dr. Jin to go to every rural place in Mississippi and Tennessee, right?” said Hawk Bo Wang, UM assistant professor of computer and information science. “This is the part we can automate.
“We can take out the routine things and focus on the challenging cases where someone really needs to talk to an expert.”
Over the next academic year, Ole Miss researchers Wang and Hyejin Park, assistant professor of communication sciences and disorders, will work with University of Memphis researchers Jin; Sohye Lee, associate professor of nursing; and Kallol Kumar Bhattacharyya, assistant professor of health systems management, to take the first step on this project.
Initially, the group will gather data to train a learning algorithm on speech patterns and other vocal cues that could indicate loneliness or depression. Research shows that people with depression and anxiety use significantly more first-person singular pronouns – such as “I” and “me” – and fewer third person pronouns such as “they,” “them,” “he” and “she.”
“Those kinds of language features could help give us indications of how they’re feeling,” Park said. “Then, in the future, we can make the AI to learn these features and then evaluate their loneliness.”
The AI-assisted autobiographies created through the project could have other uses. For patients experiencing dementia or other diseases that reduce cognitive ability, reading their life story can help jog memories, research shows.
“Think of the movie, ‘The Notebook,'” Park said. “That was a romantic story, but the process of reading their own story actually does help their cognitive function.”
While the researchers are focused on addressing social isolation, learning the stories of an older generation also has inherent value, they agreed.
“We believe AI can be a tool to help identify loneliness and depression, but I think that this is more than just research,” Lee said. “My mother-in-law always uses the old saying: ‘If an older person dies, a library has disappeared.’ It’s the same wisdom here.”
This material is based on work supported by the University of Mississippi and the University of Memphis’ Applied Artificial Intelligence Research Program.
By Clara Turnage
About the Author
