Can AI Help Prevent Homelessness? LA County’s Data-Driven Experiment
Mike Colagrossi
Predictive data and proactive outreach are transforming how Los Angeles County prevents homelessness. Partnering with the California Policy Lab at UCLA, the county is using an AI-driven model to identify residents most at risk of losing stable housing. This program, the first of its kind in the U.S., connects high-risk individuals to resources like financial assistance and housing support—often before they realize they’re in danger.
The AI model analyzes over a decade of data from health, housing, and social services to generate risk scores, which the Homelessness Prevention Unit (HPU) uses to prioritize outreach. The HPU offers tailored support, including case management and flexible cash assistance averaging $6,000 per client, addressing unique needs like transportation, mental health care, and basic household items.
Key Insights from the Report
- Equity in Action: The model performs consistently across demographics, with particular strength in identifying at-risk Black residents who are disproportionately impacted by systemic inequities.
- Promising Results: Of the 456 participants who completed the program, 86% reported living in stable housing upon discharge.
- Proactive Approach: Unlike traditional systems, the HPU identifies and contacts individuals who may not know they’re at risk, offering intervention before a crisis occurs.
“This project is a testament to how data and advanced analytics can unlock new possibilities in public services,” said Max Stevens, Chief Analytics Officer for LA County.
While challenges like low response rates (20% of those contacted enrolled in services) remain, officials emphasize that persistence is key to building trust. A randomized control trial is underway to evaluate the program’s long-term impact, with results expected in 2027.
