Gen AI in Child Welfare

Three million children were the subject of a child protection investigation in the United States last year. Four hundred thousand children spentlast night in a foster home.?

It’s important to understand some things about their experience. In addition to whatever horror brought them into foster care, they went to sleep afraid last night. They awoke desperate. They know they need to find their mom but cannot figure out how. This experience takes a toll on a little body.??

The people who serve them have tools at their disposal. First, they are well-trained, hardworking and deeply caring individuals. They have datasystems—if they are lucky, they are on a mainframe using a black screen with a green cursor. It’s more likely that they manage their work using index cards.??

I’m hard-pressed to think of a problem I could solve without the help of technology. Yesterday, I was away from home for work. I used technology to see my family, check a medical record, make a doctor’s appointment and transfer money across some joint bank accounts. I got walking directions—whispered from my wearable tech—from my hotel to a conference, including ways to navigate detours due to sidewalkclosures. I could tell you exactly where my child was, and my company, Accenture, could tell exactly where I was. If, God forbid, there had been an emergency, an Accenture BOT would have sent me instructions on how to exit the conference center, down to the level of detail of exactly which door I should use to leave.??

I honestly do not even remember how to solve problems without technology.??

And yet, we ask case workers to do it all day, every day. That seems curious insofar as they are managing such high-risk decisions.??

Their lives look like this: they receive a report that a child may be a victim of abuse or neglect. They go to the child’s physical location with little more than the words said by the individual who made the report.??

The potential of generative AI (gen AI) in these activities is extraordinary. Imagine that it is pulling information from an array of sources, harmonizing it and providing the caseworker with intel that will be critical to their decision-making. Does the child go to school every day? How many medical appointments have they missed in the last year? How many times has the family moved in the past two years? Have the adults caring for the child worked consistently for the last five years? How many times have the cops been to the house in the last two months? Are there any adults living in the home who have been named in a protective order???

The answer to these questions will not produce a decision for the case worker, but it would fuel their thinking. And, when they approach a familymore informed, it could help them put the right kids into foster care and keep the wrong kids out. This is an error we make all over the country that could be ameliorated with the use of gen AI.??

Seven paragraphs ago, I wrote about frightened children. If those comments pulled at your heartstrings, well, that just makes you a good personwith a strong heart. Those charitable feelings matter—children who are staring down child welfare systems need that from you.??

I also think we cannot have it both ways. We cannot, on one hand, say we have these charitable feelings and, on the other, know that we have failed to properly outfit caseworkers with the technology they need to do their jobs. Both things cannot be true.??

Vulnerable children and the caseworkers who serve them need our charity. They need our very best technology in equal measure.??

Learn more at www.accenture.com/ACIS or reach out to me here.?

?

About the Author?

Molly Tierney is a Managing Director and Child Welfare Lead at Accenture, where she leverages her 25+ years of expertise to enable states and cities to improve outcomes for vulnerable citizens.?