USC Researcher Targets Divorce Attorneys for Crisis Prevention

Traditional mental health interventions often fail because many individuals do not survive long enough to access clinical services. John Blosnich, an associate professor at the University of Southern California, argues that prevention must occur earlier in the process. His research identifies specific non-clinical environments where people are already navigating severe life disruptions.

Blosnich, who serves as interim co-director of the USC Center for AI in Society, refers to these environments as "industries of disruption." These include bankruptcy hearings, divorce proceedings, and self-storage counters. The core idea is that these everyday touchpoints represent critical moments when support can be connected before a crisis fully unfolds.

Identifying risk in non-clinical settings

The approach draws on social work principles of meeting people where they are. Blosnich notes that events such as divorce and bankruptcy are strongly associated with increased suicide risk. Rather than asking non-medical professionals to become therapists, the strategy focuses on recognizing signs of struggle and facilitating access to existing support systems.

Self-storage companies, for example, acknowledge that their customers often include individuals dealing with divorce, death, or displacement. By understanding these patterns, businesses in these sectors can become part of the safety net. This shifts the focus from treating symptoms after a crisis to preventing the crisis through early connection.

AI analyzes complex death data

To understand these patterns, Blosnich utilizes artificial intelligence to process data from the National Violent Death Reporting System. This CDC dataset contains hundreds of thousands of cases, combining structured fields with narrative reports from medical examiners and law enforcement. Traditional methods struggle to process the sheer volume and nuance of these narratives efficiently.

AI allows researchers to extract meaningful details from unstructured text that would be impossible to analyze manually at scale. This technology helps identify social factors that contribute to risk, such as specific types of life disruptions. The goal is to use these insights to build better early warning systems for communities.

Balancing privacy with public health

Implementing this approach requires careful navigation of privacy concerns. Using AI to analyze data from non-clinical sources raises questions about consent and data security. Blosnich emphasizes that the objective is not to surveil individuals but to understand broader social trends that impact well-being.

According to USC Today, this work highlights a trade-off between the need for granular data and the protection of individual privacy. The research aims to ensure that any data usage benefits public health outcomes without compromising the rights of those involved. This balance is essential for gaining public trust in AI-driven health initiatives.