AI and Political Freedom
The Risk to Democracy and How to Respond
Princeton University Press · 232 pages
Artificial intelligence will test democratic government as profoundly as any force in its history. Whether democracy survives that test is not a matter of fate — it is a choice, and one we still have time to make.
Also available at Amazon · Bookshop.org · Barnes & Noble
All author proceeds are donated to Protect Democracy’s AI for Democracy Action Lab.
“Essential reading for everyone, no exception.”
— Thomas Edsall, New York Times columnist
AI will remake democracy. The direction is ours to choose.
American democracy is under pressure, as institutions that we once thought stable suddenly seem precarious. At the same time, artificial intelligence has taken over much of our online life, with enormous offline consequences. In AI and Political Freedom, Matthew Botvinick examines the interaction of AI and democracy, exploring how AI is likely to challenge democracy as we move further into the twenty-first century. Botvinick — an expert on both democratic theory and AI technology — finds that the prognosis is serious. AI is poised to destabilize our democratic institutions — to attack them at their most vulnerable points and provide a useful tool for leaders with authoritarian aspirations. Yet despite the potential threat posed by the collision of AI and democracy, Botvinick argues that there are ways that we can safeguard our political freedom.
Concern about democratic backsliding emerged just as public attention turned to AI safety — but the two were seldom connected. This book makes the connection. Mass unemployment caused by technological advances creates a receptiveness to authoritarian appeals; security crises could enable governments to abuse emergency powers; and the use of AI in intelligence work creates the risk of politicized surveillance. To stop the retreat from democratic values, Botvinick argues, we must reinforce our democratic institutions and, ultimately, reimagine how democracy works.
And in that reimagining, AI itself can be an ally. Harnessed well, the same technology that strains democratic institutions can help renew them — powering forms of citizen participation richer and more engaging than any yet tried. If we play our cards right, AI will not merely be something democracy survives. It will be a source of democratic renewal.
Contents
AI, Democracy, and the New Distrust
Why the AI-safety debate and rising doubts about democratic stability emerged side by side — and why the two conversations belong together.
Foundations: Democratic Stability and the Age of AI
What actually keeps democracies stable: a framework drawn from political theory and game theory, built as a foundation for what follows.
The Age of AI, Phase I
The coming two decades: labor displacement, security crises, and AI-powered surveillance as openings for authoritarian politics.
Readying Democracy for Phase I
The near-term defense: shoring up emergency powers, surveillance law, and the other institutions AI will test first.
The Age of AI, Phase II
The remainder of the century, as AI transforms the economic and political roles on which democratic life has long rested.
Reimagining Democracy for Phase II: Ancient Athens at Scale
How AI could power forms of citizen participation richer and more direct than any yet tried — deliberative democracy at national scale.
Conclusion: The Prospects for Action
The collective-action problem at the heart of democratic reform — and the reasons for hope.
Epilogue: A Technological Republic — If You Can Keep It
The AI industry’s new pro-democracy rhetoric, examined — and what it would take for the words to be made good.
Praise for AI and Political Freedom
“Artificial intelligence has the potential to match or exceed electricity, the steam engine, and the internet as a driver of economic, cultural, and political transformation. For that reason — and much more — Matthew Botvinick’s brilliant new book, AI and Political Freedom, is essential reading for everyone, no exception.” Thomas Edsall, New York Times columnist
“This extraordinary book pulls attention away from the flashy AI distractions of the moment to the inevitable challenge AI will present to humanity and democratic governance in the next twenty years and then in the period beyond. Neither moment is hopeful, though the book brilliantly maps strategies that we must embrace now, and which could bring us to the hopeful half of this profound and terrifying future.” Lawrence Lessig, Harvard Law School
“Matthew Botvinick offers a chilling warning about the technologies he helped create: Artificial intelligence threatens the very foundations of our democracy. If we want to preserve our political freedoms, we must act now to meet its challenges.” Jack Balkin, Yale Law School
“This book tells a gripping but deeply unsettling story about democracy and technology in the twenty-first century. Drawing on a remarkable breadth of knowledge about the functioning of both AI and the machinery of state, it explains how AI is imperiling our liberty by shifting power away from citizens and towards governments. This book is essential reading for anyone who cares about how technology will shape our society, economy and politics in the years to come.” Christopher Summerfield, AI Security Institute and Oxford University
Three decades in AI. Then a turn to the law.
Matthew Botvinick has spent more than three decades at the intersection of artificial intelligence and the human mind. Trained as both a physician and a scientist — with an M.D. from Cornell University and a Ph.D. in the neural basis of cognition from Carnegie Mellon — he trained his first neural network in 1993, held faculty positions at the University of Pennsylvania and Princeton University, and went on to serve as Senior Director of Research at Google DeepMind. He is board certified in psychiatry, holds an M.A. in art history from Columbia University, and is the author of more than 150 scientific articles spanning artificial intelligence, cognitive science, and computational neuroscience.
Then he changed course. Troubled by the global democratic recession — the erosion of democratic institutions underway in many parts of the world — Botvinick turned from studying intelligence to studying the institutions that govern it. What began as work on AI policy became a full pivot: from 2023 to 2026, three decades into his scientific career, he returned to the classroom to earn a J.D. at Yale Law School, where he is now a Senior Fellow. Today he is a legal scholar as well as an AI researcher and brain scientist, and leads the AI and Rule of Law team at Anthropic.
AI and Political Freedom — the product of both careers — is his first book.
Watch: “Can AI Save Democracy?” (Kempner Institute, Harvard) · “AI as a Window on Human Intelligence” (Simons Foundation)
Occasional news about the book
A short note when the book is released, and rarely otherwise.
For press and speaking inquiries, contact James Schneider, Princeton University Press: James_Schneider@press.princeton.edu.