Each day in the United States, hundreds of thousands of people are faced with homelessness, and either live on the streets, in tents, or in crisis housing. While existing tools often enhance how we plan for and respond to homelessness, it is our understanding and relationships with the people we serve that truly drive this work.
Dr. Hsun-Ta Hsu, a professor at UNC School of Social Work, has combined AI and machine learning to help advance these efforts. After a decade of working with people on the ground, Dr. Hsu determined that an individual’s housing environment is a predictor of general well-being and long-term stability. This is especially true for those with a history of being unhoused. Dr. Hsu explained, “The neighborhoods people feel safe in have less urban decay. […] But people in Skid Row are surrounded by the environment they were homeless in. To quote them, they ‘feel homeless again.’ And living there retraumatizes them.”1 However, if individuals live in a different environment, they feel safe both within their home and their neighborhood. This finding led Dr. Hsu to begin development on a Google Maps platform that identifies urban decay — such as garbage and broken items — to predict rehousing success.
Another effort focused on using machine learning to improve the Coordinated Entry System (CES) and gain a more complete understanding of each person’s situation. Working with a team, Dr. Hsu and Eric Rice, researcher at the USC Center for Artificial Intelligence in Society, revised the CES vulnerability tool (specifically, VI-SPDAT) to prioritize which individuals would receive housing. Instead of only focusing on an individual’s vulnerabilities, the researchers included positive traits as well as general demographic data from the Homeless Management Information System (HMIS) to determine someone’s score. The Los Angeles Homeless Services Authority (LAHSA) plans to implement this revised tool in 2025 to more accurately determine who is in dire need of permanent housing placement and to reduce discrimination against specific groups.
Although tools like this optimize the decision-making process, they should not be used alone. Integrating technology and human-centered decisions can help us navigate the complexity of an individual’s experience of homelessness. As Dr. Hsu said, “Our community stakeholders, the most vulnerable population who are likely to be directly impacted by the consequences of the tool … they help us to define what the prioritization should look like.”2 Part of this is how we administer the tool. One way is through motivational interviewing, a strategy that centers on building trust and partnership in the housing process. Defined by Miller & Rollnick (2012), motivational interviewing is a “person-centered directive method of communication for enhancing intrinsic motivation to change by exploring and resolving ambivalence.”3 Motivational interviewing helps CES practitioners keep an open mind rather than make assumptions about a client’s situation – the main goal of motivational interviewing being to increase understanding, rather than to determine what is best.
Take Debra Gatlin, a participant in Rice’s work, as an example. Having been unhoused herself, she uses her experiences to help others find permanent housing. When she meets with clients, she pays close attention to facial expressions and body language. Even if they have taken the CES assessment before and received a low score, Gatlin reassesses once they open up. In one example, after reassuring her client that she was there to help, he shared new information about his military service and other challenging experiences, which increased his vulnerability score. This highlights that the way the assessment is administered can impact an individual’s housing opportunities. Rather than assuming that the client’s past results define their options, Gatlin takes the time to listen and establish trust, which ultimately helps to match her clients with more appropriate housing options.
Regardless of which tools and technology we have, problems can quickly arise when we do not understand the people we are working to help. Motivational interviewing can open lines of communication, allowing the client to explain their background, explore their reasons for change, and feel more confident about their decisions. Most importantly, it ensures that an individual’s needs and values are at the heart of the process.
References:
-
- UNC Endeavors, “Home Truths,” UNC Endeavors, accessed March 10, 2025, https://endeavors.unc.edu/home-truths/.
- Carly Stern, “LA thinks AI could help decide which homeless people get scarce housing — and which don’t,” Vox, accessed March 10, 2025, https://www.vox.com
- Elizabeth Jenkins, Ph.D. “Motivational Interviewing,” National Center on Homelessness Among Veterans, accessed March 10, 2025, https://www.va.gov/HOMELESS/nchav/docs/3m_MI.pdf.