ReDDDot Phase 1: Planning Grant: Prioritization of Housing & Behavioral Health Services to Individuals and Families

Homelessness is a critical social issue in the United States, exacerbated by the complex interplay of behavioral health problems such as mental health disorders and substance abuse. This project addresses this issue by enhancing the Allegheny County Department of Human Services' (DHS) existing Allegheny Housing Assessment (AHA) tool, which prioritizes housing services for homeless individuals based on risk scores (or assessments). The primary goal of the project is to develop AHA+, an advanced version of the current tool that integrates sophisticated artificial intelligence (AI) to improve transparency and decision-making. By providing clear explanations for risk scores and optimizing service allocation, AHA+ aims to foster trust and efficiency in the delivery of housing and behavioral health services. This initiative not only aligns with the National Science Foundation's (NSF) mission to advance knowledge but also sets a precedent for the responsible use of AI in social services, potentially transforming public perceptions and applications of AI in various fields.The project focuses on two main research thrusts. The first thrust involves developing and implementing a range of automated explanation systems for the current AHA tool to enhance transparency. These systems will provide varied types of explanations, such as rule-based, example-based, and counterfactual explanations, to cater to different stakeholders' needs. User studies will be conducted with DHS staff and homeless clients to refine these systems, ensuring they are user-centric and effective. The second thrust aims to optimize the allocation of housing and behavioral health services using advanced optimization algorithms. This includes developing AI based predict-then-optimize style optimization algorithms and models that recommend the best combination of services for each client based on their specific needs, moving beyond the current ad-hoc decision-making process. By continuously engaging stakeholders in the design and evaluation of these systems, the project ensures that the AI technologies developed are both practically viable and ethically sound. Through this participatory approach, the project seeks to create a more effective and equitable service delivery model, ultimately improving outcomes for homeless individuals and families.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.