Insights on AI governance, public sector innovation, and democracy
News That Caught Our Eye
What determines whether AI works, after the policy is written.
Making government services easier to use doesn’t start with a website redesign. It starts by talking to residents, choosing the right first project, and creating the conditions for UX to succeed. Lessons from U.S. Digital Response’s three-part InnovateUS series show how agencies can build user-centered services that deliver measurable results.
Making government services easier to use doesn’t start with a website redesign. It starts by talking to residents, choosing the right first project, and creating the conditions for UX to succeed. Lessons from U.S. Digital Response’s three-part InnovateUS series show how agencies can build user-centered services that deliver measurable results.
New research from the AI Risk Explorer SPAR project examines how AI is reshaping conflict through cloud infrastructure, commercial technology companies, and the digital systems that now underpin both military operations and democratic governance.
New research from the AI Risk Explorer SPAR project examines how AI is reshaping conflict through cloud infrastructure, commercial technology companies, and the digital systems that now underpin both military operations and democratic governance.
Mississippi has released its statewide AI framework. The harder challenge is turning guidance into practice. Drawing on firsthand experience leading the state’s AI workforce efforts, Dr. Kollin Napier of MAIN examines why successful AI adoption depends on sustained investment in training, governance, staffing, and implementation.
Mississippi has released its statewide AI framework. The harder challenge is turning guidance into practice. Drawing on firsthand experience leading the state’s AI workforce efforts, Dr. Kollin Napier of MAIN examines why successful AI adoption depends on sustained investment in training, governance, staffing, and implementation.
A weekly roundup of the most intriguing and impactful stories from the world of artificial intelligence
Mississippi has a statewide AI framework, but Dr. Kollin Napier argues that the real test starts the next day. In a new Reboot Democracy piece, he reflects on what it takes to move from published policy to actual adoption: sustained investment in workforce training, governance, staffing, and implementation capacity. From the AI Risk Explorer SPAR project, Elana Banin and colleagues argue that AI is pulling commercial tech firms into military operations and making cloud infrastructure, data centers, subsea cables, and AI models strategic targets. And a new piece from U.S. Digital Response and InnovateUS outlines what it takes to build UX capacity in government: not more technology, but leadership, cross-agency collaboration, and starting with residents
Making government services easier to use doesn’t start with a website redesign. It starts by talking to residents, choosing the right first project, and creating the conditions for UX to succeed. Lessons from U.S. Digital Response’s three-part InnovateUS series show how agencies can build user-centered services that deliver measurable results.
New research from the AI Risk Explorer SPAR project examines how AI is reshaping conflict through cloud infrastructure, commercial technology companies, and the digital systems that now underpin both military operations and democratic governance.
Mississippi has released its statewide AI framework. The harder challenge is turning guidance into practice. Drawing on firsthand experience leading the state’s AI workforce efforts, Dr. Kollin Napier of MAIN examines why successful AI adoption depends on sustained investment in training, governance, staffing, and implementation.
Available now in print and in audiobook! "Reboot: AI and The Race to Save Democracy" by Beth Simone Noveck
AI adoption in government is accelerating. The next challenge is making sure it creates public value. InnovateUS’s Fall 2026 Live Learning Series explores how public professionals can move beyond using AI to measuring its impact, governing it responsibly, and ensuring it improves outcomes for the people and communities government serves.
This excerpt from Reboot: AI and the Race to Save Democracy was originally published in Fast Company on June 25, 2026.
Amazon, Microsoft, and Google control 85% of Canada's cloud market. In a new Reboot Democracy piece, David Eaves and Curtis McCord argue this concentration is a democracy problem, not just a competition one, because public institutions lose leverage over the infrastructure they depend on. The GovLab and InnovateUS launch an AI Coach for Public Engagement that helps public servants move from training to real project planning. And in Part II of their Habermolt response, José Luis Martí and Beth Simone Noveck argue that delegating political judgment to AI agents risks replacing democratic self-government with algorithmic authority.
A new report from the Canadian Anti-Monopoly Project argues that cloud sovereignty is about more than building domestic alternatives. It is about ensuring governments, businesses, and public institutions have autonomy and agency. Using Canada’s highly concentrated cloud market as a case study, the authors explore how procurement policy, competition rules, and interoperability standards could reduce dependence on hyperscalers and create a more competitive marketplace for compute.
In Part II, we examine the project’s core claim that AI agents could represent citizens in democratic deliberation. We argue that representation requires more than predicting preferences, that democratic participation is essential to civic learning and self-government, and that increasingly capable AI agents risk shifting authority away from citizens and toward algorithms. Democracy depends on people governing together, not on delegating judgment to machines.
A new AI for Impact-engineered public engagement coach is helping public servants turn what they learned about engagement into practical action. Built by The GovLab, the tool guides practitioners through real planning decisions, asks targeted questions about their specific project, and draws on engagement frameworks, case studies, and practitioner expertise to help them design more effective public participation processes.
Democracy needs a moonshot — and Beth Noveck's new book offers the blueprint. Launching next week from Yale University Press, Reboot: AI and the Race to Save Democracy makes the case that AI can strengthen democratic institutions when designed for public purpose, governed responsibly, and built to help governments listen, decide, and act more effectively. This special edition gathers the pre-launch conversation: Brookings TechTank with Darrell West, the Tavis Smiley Show, and Rick Flynn Presents, with feature interviews in Route Fifty and The Herald — alongside her new Reboot Democracy essay with José L. Martí, which critiques Habermolt, an experimental project that uses AI agents to deliberate and vote. Elana Banin sits down with Anya Kamenetz to ask why families have been largely absent from AI policy debates, and the Barcelona Public-Interest AI Accelerator team reflects on what it takes to train the next generation of AI-and-democracy leaders.
A new accelerator in Barcelona is exploring what it takes to prepare future leaders to navigate the emerging field of public-interest AI. The experience underscored that moving from principles to practice demands leaders who understand the technical, political, economic, and ethical dimensions of AI. Leaders who are equipped to ensure these technologies serve public goals, protect rights, and respond to the needs of diverse communities.
Beth Simone Noveck and José Marti critique Habermolt, an experimental project that uses AI agents to deliberate and vote on behalf of people. They argue that the project reflects a longstanding strain of democratic theory that views public participation as a problem rather than a resource. At a moment of declining trust and growing institutional challenges, there should be less focus on whether AI can simulate democracy and more on how it can help governments listen, learn, and solve problems with the public. This Essay is Part 1 of 2 by Noveck and Martí.
Beth Simone Noveck recently appeared on the Brookings TechTank podcast to discuss her forthcoming book, Reboot: AI and the Race to Save Democracy (Yale University Press), arguing that AI offers powerful opportunities to strengthen democratic institutions and making the case for a "possibilist" approach to building tools that make democracy more responsive and effective.
InnovateUS 