The city of Los Angeles is implementing an AI pilot program to forecast homelessness and distribute assistance.

• In Los Angeles, a pilot program utilizes predictive AI to identify individuals and families at risk of homelessness, providing aid to help them stay housed.
• Nearly 800 individuals and families at risk of homelessness have been served by the technology launched in 2021, with 86% of participants retaining permanent housing after leaving the program, as stated by Dana Vanderford, an associate director at the county's Department of Health Services.
• She stated that individuals and families can access between $4,000 and $8,000.

Last December, Courtney Peterson, a single mom, was laid off from her job at an inpatient transitional living program, which had allowed her to bring her seven-year-old son to work and covered their rent in a Van Nuys neighborhood apartment complex in Los Angeles, where they had resided for 1.5 years.

Peterson stated that she started investigating possible solutions for assistance, as she was worried about paying January's rent. She mentioned that when her son was a baby, they resided in a travel trailer, which she did not wish to revert to.

"Peterson explained to CNBC that he tried contacting local churches and organizations that claimed to provide rent assistance. However, many of them required him to have active eviction notices before they could offer help. Feeling desperate, Peterson realized he had exhausted all his options after reaching out to everyone he could think of with no success."

Instead of an eviction notice, Peterson received a letter from the Homelessness Prevention Unit within the Los Angeles County Department of Health Services, offering a lifeline through a pilot program that uses predictive artificial intelligence to identify individuals and families at risk of becoming homeless and provides aid to help them stabilize and remain housed.

In 2023, California's homeless population increased by more than 30% since 2007, with over 181,000 individuals, according to data from the U.S Department of Housing and Urban Development. Meanwhile, a report from the Auditor of the State of California revealed that the state spent a staggering $24 billion on addressing homelessness from 2018 to 2023.

Since its launch in 2021, the technology has assisted the department in serving over 790 individuals and families at risk of homelessness, with 86% of participants maintaining permanent housing after exiting the program, as stated by Dana Vanderford, the associate director of homelessness prevention at the county's Department of Health Services.

The American Rescue Plan Act provides between $4,000 and $8,000 in funding for individuals and families, with the majority of the funds allocated for the program. However, reaching out to individuals and convincing them that the offer is legitimate and not a scam can be difficult. Once contact is made, aid is swiftly provided.

"Vanderford stated that meeting clients within days of a loss of housing or after a medical emergency is crucial. Their ability to provide resources and prevent imminent loss of housing for 86 percent of their clients is remarkable."

Peterson and her son were given $8,000 to cover their rent, utilities, and basic needs, enabling her to remain in her apartment while she searches for a new job. The program supports clients for four months and then checks in with them at the six-month, 12-month, and 18-month marks after discharge. Case workers, such as Amber Lung, who assisted Peterson, can attest to the significance of preventative work through their own experiences.

If people lose their housing, it can be challenging to regain it, and if they end up in a shelter or on the streets, it can be even more difficult. However, if we can provide a little support to help them retain their housing, it can be much easier to stabilize their situation.

Predicting Risk

Over several years, the California Policy Lab at UCLA developed an AI model using data from Los Angeles County's Chief Information Office. The CIO integrated data from seven different county departments, including emergency room visits, behavioral health care, and large public benefits programs from food stamps to income support and homeless services. The program also pulled data from the criminal justice system, according to Janey Rountree, executive director of the California Policy Lab.

The model's accuracy was tested during a period when the policy lab had data linked together over many years to make predictions about who would experience homelessness.

The lab created an anonymized list of individuals ranked from highest risk to lowest based on patterns identified in who experienced homelessness, which the county used to reach out to people who may be at risk of losing housing before it happened.

A study on data privacy using 1990 U.S. Census data found that 87% of Americans could be identified by using ZIP code, birth date, and gender.

"The root cause of homelessness in California is a structural economic issue, not individual risk factors, as Rountree stated."

The Policy Lab gave the software to the county without charging, and has no plans to profit from it, Rountree stated. By collaborating closely with individuals who possess specialized knowledge in various fields, such as educators and social workers, AI can contribute to enhancing positive social outcomes, she added.

""Emphasizing the importance of testing and innovating around prevention for communities experiencing homelessness is crucial, as it is a relatively new strategy in the lifespan of homeless services. We need more evidence and experiments to find people at risk. This is just one way to do that," she said."

In 2017, the National Alliance to End Homelessness discovered that the average yearly cost to the taxpayer for a chronically homeless individual is $35,578. However, when these individuals are placed in supportive housing, the costs are reduced by nearly half.

Santa Clara County and San Diego County are also considering a similar program to Los Angeles County, as initial conversations have been held, Vanderford stated.

Government Use of Artificial Intelligence

Government agencies' use of AI has been scrutinized due to potential consequences, including wrongful arrests resulting from police reliance on AI technology. In California, voters rejected a plan to replace the state's bail system with an algorithm to determine individual risk in 2020, citing concerns about bias in the justice system.

According to Margaret Mitchell, the chief ethics scientist at AI startup Hugging Face, the context of use and safety of identifiable information are crucial factors in determining the ethics of government use of AI, even if the information is anonymized. Additionally, Mitchell emphasizes the importance of obtaining informed consent from individuals seeking assistance from government programs.

"Mitchell pointed out that the people may not be fully aware of the signals being collected and the potential risks associated with it, as well as the dual use concerns for malicious purposes against them. Additionally, there is the issue of how long this data is being stored and who may eventually have access to it."

From a utilitarian perspective, while the technology aims to aid those in need before their housing is lost in Los Angeles County, there are broader questions to consider.

She raised concerns such as 'What is the cost to the taxpayer and how effective is this system in preventing homelessness?'

Peterson is currently searching for remote work while also aiming to become a licensed vocational nurse and eventually purchase a home for her son.

"The program's aid has been significant because my son and I have not always had stability. Knowing that we can call this place home and not have to worry about moving or finding new friends has meant a lot to both of us."