The team explored using LLMs to interpret the Program Operations Manual System (POMS) into plain language logic models and flowcharts as educational resources for SSI and SSDI eligibility, benchmarking LLMs in RAG methods for reliability in answering queries and providing useful instructions to users.
Errors in administrative processes are costly and burdensome for clients but are understudied. Using U.S. Unemployment Insurance data, this study finds that while automation improves accuracy in simpler programs, it can increase errors in more complex ones.
BenCon 2024 explored state and federal AI governance, highlighting the rapid increase in AI-related legislation and executive orders. Panelists emphasized the importance of experimentation, learning, and collaboration between government levels, teams, agencies, and external partners.
This report investigates how D.C. government agencies use automated decision-making (ADM) systems and highlights their risks to privacy, fairness, and accountability in public services.
These principles and best practices for AI developers and employers to center the well-being of workers in the development and deployment of AI in the workplace and to value workers as the essential resources they are.
The study investigates how state agencies administering SNAP comply with Title VI of the Civil Rights Act by providing language access for individuals with limited English proficiency (LEP).
For the past year, modernization teams at the Department of Labor (DOL) have been helping states identify opportunities to automate rote, non-discretionary, manual tasks, with the goal of helping them speed up the time that it takes to process claims. This post provides more context on Robotic Process Automation (RPA) and potential use cases in unemployment insurance.
An in-depth report that examines how states use automated eligibility algorithms for home and community-based services (HCBS) under Medicaid and assesses their implications for access and fairness.
The team developed an AI solution to assist benefit navigators with in-the-moment program information, finding that while LLMs are useful for summarizing and interpreting text, they are not ideal for implementing strict formulas like benefit calculations, but can accelerate the eligibility process by leveraging their strengths in general tasks.
