Insights on AI, Governance and Democracy
The Air Force is quietly auctioning off slices of its bases for private AI data centers. They call it innovation; it looks like privatization. Fifty-year leases, 3,000 acres of military land, and no public say. If this is how we build the future, who’s really in command?
When Brazil's federal government launched its 2023 Participatory Pluriannual Plan, the response was massive: 1.4 million participants submitted over 8,200 proposals through the Brasil Participativo platform. But volume creates a challenge, as manually processing thousands of contributions is slow and resource-intensive, often causing valuable insights to slip through the cracks. Now, Brazil is pioneering an open-source AI system that automatically analyzes citizen feedback, generates comprehensive reports, and tracks which suggestions made it into final policies. The result is a new model for democratic intelligence, one that transforms the flood of public input into structured, actionable knowledge without losing the nuance of individual voices.
The Agentic State is an ambitious and inspiring blueprint for rebuilding government around AI agents that can act and decide autonomously. It powerfully diagnoses real failures in how the public sector designs, delivers, and manages services. While AI is giving us ways to accelerate change, the prescribed cure may be premature: most of what’s broken in government requires organizational reform, not automation.
The danger isn’t that AI will make us dumber—it’s that governments, companies, and schools won’t make us smarter with it. As policymakers stall and corporations automate, the burden of using AI wisely now falls on us. Like shoppers at the fish counter, we need to learn to read the labels—to know what’s safe, what’s risky, and how to choose well.
Public service professionals from across the globe have come together to learn from one another how to use AI to improve governance. From St. Louis cutting hiring times from 12 to 2 months, Hamburg analyzing 11,000 public comments in days, and New Jersey reducing Spanish-language form completion from 4 hours to 25 minutes, it becomes clear that AI works for democracy when we build the foundation first.
As headlines warn of “superintelligent AI” threatening human extinction, a new open letter reignites familiar fears. But beneath the apocalyptic rhetoric lies a deeper problem. The narrative around artificial superintelligence, long embraced by Big Tech, diverts attention from the real and immediate challenges of AI and how our democratic institutions can address them.
After two years of research, the RethinkAI collaborative released Making AI Work for the Public—a comprehensive field review of how U.S. governments adopt AI. Since 2019, over 1,600 AI-related bills have been introduced, but most focus on guardrails, not proactive strategy. Meanwhile, cities are piloting translation tools, engagement platforms, and predictive systems, often led by Chief Information Officers, taking on new strategic roles. The report challenges civic tech’s efficiency-first legacy and proposes a new governance model—ALT: Adapt to anticipate needs, Listen to understand communities, and build Trust through two-way accountability.
Officials in Hamburg had long struggled with the fact that while citizens submitted thousands of comments on planning projects, only a fraction could realistically be read and processed. Making sense of feedback from a single engagement could once occupy five full-time employees for more than a week and chill any desire to do a follow-up conversation. Learn about how Hamburg built its own open source artificial intelligence to make sense of citizen feedback on a scale and speed that was once unimaginable.
Boston became a living laboratory for democratic innovation last week, as two major convenings—the Civic AI Summit at Northeastern and Harvard’s Digital Democracy showcase—brought together leaders reshaping how technology serves the public good. From new tools that open up lawmaking and procurement to partnerships that align city and state AI strategies, Boston’s approach offers a model nationally for how AI can strengthen democracy through human-centered design, transparency, and collaboration.
As states take center stage in shaping how the U.S. adapts to artificial intelligence, their choices will determine not just whether America keeps pace, but whether it thrives. This summer, Princeton University’s Center for Information Technology Policy convened state AI officers, researchers, entrepreneurs, and technologists for “Shaping the Future of AI: A National Gathering for State AI Leaders.” The two-day working conference focused on building practical, responsible frameworks for public-sector AI implementation. New America CEO Anne-Marie Slaughter closed the convening with a wide-ranging keynote that called for public AI infrastructure, trust-based governance, and co-creation across sectors. What follows is a summary of her 10 core takeaways.
New York City has thousands of parks, plazas, and public courtyards, but no easy way to find them. Using “vibe coding,” open data, and generative AI, one civic technologist built a map of every public space in the five boroughs. This is the story of NYC Public Space, an app that stitches together fragmented government datasets, AI-generated descriptions, and community-sourced updates to make the city’s public realm more visible and usable. It’s also a case study in how AI can help public interest technologists move faster, build smarter, and turn open data into real public value.
In late September, the UN adopted a global AI resolution backed by all 193 Member States, a diplomatic milestone. But one that risks repeating old patterns of top-down governance. The new Reboot Democracy Blog Editor, Elana Banin, argues that legitimacy doesn’t come from declarations, but from grounded, democratic practice. From California to Vietnam, she explores what real AI governance looks like and lays out three strategic tests the UN must pass to matter.
Large Language Models are becoming common tools in the communications toolkit, but not all uses are created equal. In this new post from the AIMES Lab at Northeastern University, John Wihbey and Samantha D’Alonzo offer research-backed guidance on when to use LLMs in the survey pipeline and when to steer clear. The research indicates that AI is a powerful assistant for refining survey questions and testing hypotheses, but a poor substitute for actual human respondents. Drawing on more than 30 academic studies, this piece lays out a practical, hybrid approach to “silicon sampling” that helps practitioners strengthen research integrity without falling for AI’s easy shortcuts.
AI’s biggest breakthroughs were built on public datasets; the next wave should be, too. If governments make data AI-ready and keep access open, we can power tools that explain laws, improve services, and widen participation. In return, companies that train on taxpayer data should give back, through open licenses, benchmarks, and capacity that strengthen public institutions.
Law professor and political commentator José Luis Martí argues that Spain’s Supreme Court ruling on Bosco is a democratic milestone—establishing algorithmic transparency as a constitutional principle and putting AI at the service of democracy.