Social Workers and AI: How New Voice Tech Can Save Lives

Article Featured Image

Technology has the power to fundamentally change how we live for the better. This has proven evident in the field of social work, one of the most demanding professions in America, and often one of the most understaffed and underfunded. Especially in urban centers like Chicago, New York, and Los Angeles, there are often hundreds of unfilled vacancies for social workers, both in adult and child services. The result is a child-to-social worker ratio of up to 1.000 to one (when a ratio of 50 to one is recommended).

Artificial intelligence (AI) is helping to close that gap, providing much-needed clerical and observational support that can drastically amplify the abilities of even one social worker in the field. From passive capture of emotional data on calls to augmentation of paperwork, social workers are getting a significant boost from advanced technology in the field. AI might be the long-awaited solution to support social workers in providing the best possible service and care.

Reducing the Human Factor in the Field

Social workers are tasked with working closely with sometimes hundreds of high-risk individuals. They evaluate potentially abusive home situations for children, address suicide risk, and have to make harrowing recommendations based on their observations. But they are only human, and with so many patients, they can easily miss issues

This was tragically on display in the recent Netflix documentary, Trials of Gabriel Fernandez. The film looked at the failure of social services to identify the risk Gabriel's mother posed to him. She repeatedly lied on calls meant to identify base potential risks to the child, and her family knew of her history of dishonesty, but the social worker, who interacted with dozens of people every day and had just met her, was not given enough evidence or indicators to trigger suspicion.

And yet, the documentary highlights how AI could have identified the risk based on age, history of calls, history of mental illness, the existing criminal records of Gabriel's mother and her boyfriend, and more. By removing human bias and evaluating cases based on data, a life might have been saved.

Similarly, social workers could be supported by AI in identifying potential suicide risks and intervening appropriately

Social workers have traditionally been left to rely on behavioral history (depression, stress levels, and substance abuse rates), and the accuracy of these predictions is middling at best. Machine learning algorithms can evaluate thousands of medical records and build a model to predict higher suicide rates. Similar tools in development at the University of Southern California are being used to identify the risk of drug use relapse. The result could be groundbreaking, helping both clinicians and social workers identify high-risk individuals and provide added care and check-in more often. The U.S. military and the Veterans Administration are already researching how best to implement the technology.

Streamlining Processes to Free Valuable Time

AI has the power to remove much of the added burden placed on these individuals by streamlining processing and paperwork and freeing more time to focus on the most important things a social worker does every day. AI has been used effectively in the medical field to help reduce the percentage of time doctors spend on paperwork and administrative work.

Social workers who routinely spend 60-70 hours a week processing paperwork and following up with the people in their portfolio just to make sure every phone call is made each week can shift their resources to the tasks that have the greatest impact - spending more time on the phone, making in-person visits, sitting down to talk with people who need the support.

The Importance of Emotion in AI for Social Work

But in dealing with human beings and emotional problems, AI needs to be sensitive to the nuances of human behavior to fully support social workers in the field. Emotional recognition in AI enables this by evaluating dozens of factors in a human voice to build a complete profile of someone.

Because social work follow-ups are so often done by phone, it is possible to integrate emotion AI technology into these systems and evaluate the responses in real time. By capturing user-generated data and comparing reactions to certain stimuli, machine learning algorithms can recognize human behavioral patterns with a high degree of accuracy. The result is a better ability to identify when someone is lying, respond to a rise in anger, raise a flag if someone poses a suicide or drug use relapse risk, and more. For social workers who are often overworked and interact with so many people every week, this kind of real-time insight can be invaluable to prioritize cases, intervene more accurately, and make the right decisions in high-pressure situations.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues