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.
In May 2020, Stanford's HAI hosted a workshop to discuss the performance of facial recognition technologies that included leading computer scientists, legal scholars, and representatives from industry, government, and civil society. The white paper this workshop produced seeks to answer key questions in improving understandings of this rapidly changing space.
The team developed an AI-powered explanation feature that effectively translates complex, multi-program policy calculations into clear and accessible explanations, enabling users to explore "what-if" scenarios and understand key factors influencing benefit amounts and eligibility thresholds.
This guide, directed at poverty lawyers, explains automated decision-making systems so lawyers and advocates can better identify the source of their clients' problems and advocate on their behalf. Relevant for practitioners, this report covers key questions around automated decision-making systems.
This reporting explores how algorithms used to screen prospective tenants, including those waiting for public housing, can block renters from housing based on faulty information.
In this piece, the Digital Benefits Network shares several sources—from journalistic pieces, to reports and academic articles—we’ve found useful and interesting in our reading on automation and artificial intelligence.
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.
This academic article develops a framework for evaluating whether and how automated decision-making welfare systems introduce new harms and burdens for claimants, focusing on an example case from Germany.
This policy brief offers recommendations to policymakers relating to the computational and human sides of facial recognition technologies based on a May 2020 workshop with leading computer scientists, legal scholars, and representatives from industry, government, and civil society
The team explored the performance of various AI chatbots and LLMs in supporting the adoption of Rules as Code for SNAP and Medicaid policies using policy data from Georgia and Oklahoma.
The team developed an application to simplify Medicaid and CHIP applications through LLM APIs while addressing limitations such as hallucinations and outdated information by implementing a selective input process for clean and current data.
