How should AI be deployed in welfare, taxation, policing, courts, immigration, intelligence, and public administration?
Agency adoption reached far further by 2020 than most realized — mostly unglamorous, mostly ungoverned. The live split is between procedural-safeguard approaches (due process, contestability) and use-case prohibitions, while adoption outpaces both.
View on the map → · Open in Browse →
Adoption ran ahead of governance: 3.4 million federal users can now reach frontier chatbots through GSA's OneGov deals, and the first major ruling held that an agency's chatbot-driven classifications are the government's own act for constitutional purposes — outsourcing a decision to a model does not launder state action. The deeper problem now has a research agenda: automating the executive branch removes the human friction that checks presidential power.
DOGE selected the AI tool, formulated the prompt, and defined the operative viewpoint-based criterion.
Analyzes the first major ruling holding an agency's chatbot-driven classifications to be 'the Government's own for constitutional purposes' — a landmark for the procedural-safeguard side of the government-AI debate.
Sooner or later, in this administration or a future one, AI will come for the federal workforce.
Frames the constitutional problem beneath agency AI adoption: automating the executive branch removes the human friction that currently checks presidential power, and proposes rule-of-law safeguards before that happens.
More than 120 orders have been placed against OneGov's AI offerings
The best single measurement of how far frontier chatbots have diffused into the US federal workforce — 3.4 million users — showing adoption running far ahead of governance frameworks.
Accuracy mattered more than anything else we tested
Survey evidence on public preferences over government AI: citizens' tolerance varies sharply by use case, with policing and courts demanding the most accuracy, explainability, and human oversight — empirical grounding for use-case-tiered rules.