AI Agents in Government and policy
AI agents in government and public services, and the laws and policies on AI agents: what happened, and what it means for public leaders. Strictly nonpartisan.
Patterns for public leaders
- Steps Only a Person May Take
Name the decisions no agent may take, such as a benefit or a penalty; agents prepare them, and a person takes them. - Log Actions, Not ConversationsRecord every action an agent takes: who, what, who approved it, and how to undo it.
- No Wrong DoorPeople reach a public service through whatever channel they use, and every channel draws on the same rules and tells them where their request landed.
- Living Versus FrozenDecide which documents may change and which must not; an adopted rule changes only by a new, approved version.
- Check at the DoorOne review sits at the point of publishing: nothing goes out until the exact page or message has passed it.
AI agents in government and policy, explained
What does “AI governance” mean?
AI governance is the set of rules, roles, and checks that decide how AI is built and used, who answers for it, and how problems get caught and fixed. More on AI governance
What does “AI agent” mean?
Software that uses an AI model to decide its own steps and use tools to finish a task for someone. More on AI agent
What does “Human in the loop” mean?
A person who approves an AI's work at some step. More on Human in the loop
What does “AI safety” mean?
AI safety is the work of making sure AI systems do not cause harm, whether by accident, through misuse, or by acting against the wishes of the people responsible for them. More on AI safety
What does “Guardrails” mean?
Checks that stop an agent from doing what it should not, built into the system rather than left to the agent's good behavior. More on Guardrails
What does “Foundation model” mean?
A foundation model is a large AI model trained on huge amounts of general data so that many different products, from chat assistants to AI agents, can be built on top of it. More on Foundation model
Our research
On-ramps: How Requests Find Their Process in an Agentic Organization
Our research note on routing each request to the process that handles it. It describes an early test at one small nonprofit, and is not specific to government.
Strictly nonpartisan: GSSF is a 501(c)(3) foundation, and we do not support or oppose any candidate, party, or ballot measure. This section reports on laws and policies; we do not lobby or urge readers to contact officials about legislation. Nothing here is legal, financial, investment, or medical advice.
Events in Government
A research conference on how people and AI agents work together, themed "From Interaction to Agency: Navigating Autonomy in HAI."A yearly conference on digital learning for schools, universities and workplaces. This year's theme is trust, agency and how learning can prepare people to run and oversee intelligent systems responsibly.A program for people who build, buy, and regulate AI agents in companies and government, covering how to coordinate many agents and keep them secure; the official page does not list ticket prices.A gathering of government and international leaders on the theme "Shaping Future Governments," with the future of work and education on the agenda; the 2027 page does not yet mention AI sessions.A research conference at Helsinki Congress Paasitorni where AI and human-computer interaction meet, focused on practical problems in how people work with intelligent systems. marks our own words: our view, our summaries, and our book. About us · Our research · Our standards · Corrections · Report an error · Write for us · Press · Privacy · TermsPublished by GSSF Press, a program of the Global Student Startup Foundation · 1837 McFarlane Street, San Marino, CA 91108 · [email protected] · WhatsApp +1 213-222-8044