This report analyzes the growing use of generative AI, particularly large language models, in enabling and scaling fraudulent activities, exploring the evolving tactics, risks, and potential countermeasures.
The Public Design Evidence Review examines how design practices can improve public policies and services across the UK, exploring what good “public design” looks like, how it’s being used, and what enables or inhibits its impact.
This report examines how governments use AI systems to allocate public resources and provides recommendations to ensure these tools promote equity, transparency, and fairness.
This report provides supplemental estimates on how Public Law 119-21—tied to H.R. 1—will affect SNAP participation, benefits, and state administrative costs over 2025–2034.
This framework provides a structured approach for ensuring responsible and transparent use of AI systems across government, emphasizing governance, data integrity, performance evaluation, and continuous monitoring.
A practical guide for advocates that explains how automated benefit notices are generated, where common notice failures originate, and how to push for effective fixes.
This publication summarizes a body of research about how state benefits administering agencies build and maintain integrated eligibility and enrollment (IEE) systems. It is an easy to reference guide for state administrators, legislators, advocates, and delivery partners.
A research brief explaining how work requirements in programs like Medicaid and SNAP reduce coverage, increase administrative costs, and push eligible people deeper into poverty without improving employment outcomes.
This section of the Building Resilience plan outlines strategies to improve the long-term solvency and sustainability of state unemployment insurance (UI) trust funds through better funding practices and legislative reform.
This report explores policy options Utah and other states can adopt to mitigate benefit cliffs, which occur when small income increases lead to sudden loss of public assistance.
