Chapin Researchers Create AI-assisted Prototype for Systematic Reviews

Keeping youth homelessness evidence current to impact policy, practice

Youth homelessness is a global issue with researchers worldwide focusing on causes and solutions and publishing their studies and results. The research is ongoing and prolific and can be overwhelming to track, but it’s essential for those studying the issue to stay current and to have a comprehensive overview of relevant research.

Living Evidence for Youth Homelessness, a new public research website project, brings together complementary approaches led by Chapin Hall researchers. Chapin Hall Senior Researcher Dr. Kay Chansiri led the project’s AI-assisted approach and the website development, exploring how AI tools can support a “living evidence” approach to youth homelessness intervention research. Chapin Hall Researcher Dr. Sara Semborski led the traditional, human-driven research and systematic review .

The Living Evidence for Youth Homelessness project focuses on the challenge of synthesizing evidence so that it can inform policy and practice. In fields like youth homelessness, where communities, service providers, policymakers, and researchers need timely evidence to address urgent situations, the gap between available research and current decision-making can be especially important.

The website presents a prototype for using AI-assisted screening to help make systematic review updates more efficient, transparent, and repeatable. The prototype focuses on screening published titles and abstracts, one of the most time-intensive stages of systematic review work. AI-assisted screening outputs were not treated as final inclusion decisions, but were compared with human reviewer decisions to evaluate how well the workflow supported screening while keeping human judgment central. For each data phase, human validation involved two human review teams.

In the second phase, the AI prompt caught 91% of the studies human reviewers considered relevant. This is a strong result for early abstract screening, where the main concern is avoiding missed studies; a relevant study excluded too early may never reach full-text review. The results suggest AI-assisted screening can support a living systematic review when used as a human-in-the-loop triage process.

The website also includes Living Notes, a shared space for youth homelessness researchers, service providers, advocates, and community partners to follow emerging evidence and connect research findings to policy and practice conversations. The first Living Note highlights findings from a systematic review of youth homelessness intervention effectiveness studies. Future notes may include updates from newer studies, under-review work, evidence gaps, practice questions, and emerging issues from the field.

Together, the website and related resources demonstrate how AI can support living evidence infrastructure for youth homelessness research, policy, and practice. It shows how AI-assisted screening can improve efficiency and transparency in review updating while still relying on human reviewers for validation and final decisions. The project invites conversation about how AI can be used responsibly to strengthen the evidence base on youth homelessness, and it offers a model that could be replicated in other fields of study.