Book Review: AI and Political Freedom — Why Democracy's Weak Points Matter More Than Deepfakes
Matthew Botvinick's new book from Princeton University Press argues that the real threat AI poses to democracy is not the fake video but the pressure it will place on the institutions that keep power in check. A review of its method and what leaders should take from it.
Executive summary
- The book: AI and Political Freedom: The Risk to Democracy and How to Respond, by Matthew Botvinick (Princeton University Press, published October 6, 2026).
- The thesis: AI endangers democracy less through misinformation than through the stress it puts on democracy's structural weak points: employment, emergency powers and surveillance.
- The method: a framework from political theory and game theory on what keeps democracies stable, applied to two time horizons.
- Why it matters: it moves the conversation from content moderation to institutional design, the place where governments, parties and companies actually make decisions.
The argument
Most public debate about AI and politics has focused on what we can see: the deepfake of a candidate, the bot network, the synthetic voice on a robocall. Botvinick, a former Senior Director of Research at Google DeepMind and now a Senior Fellow at Yale Law School, asks a different question. What happens to the conditions that make democracy stable when a technology this powerful arrives?
His answer is organized in two phases. In the first, roughly the next two decades, he identifies three openings for authoritarian advantage:
- Labor displacement, which can make large groups of citizens more receptive to authoritarian appeals.
- Security crises, which give governments a pretext to stretch emergency powers.
- AI-enabled surveillance, which lowers the cost of monitoring opponents and journalists.
In the second phase, over the rest of the century, he argues that AI may change the economic and political roles that democratic life has always relied on. The book closes with a constructive proposal: using AI to make deliberation possible at national scale, what he calls "ancient Athens at scale."
The methodological value
The book's strength is its frame. Instead of cataloguing harms, Botvinick starts from a model of democratic stability, the line of work associated with Barry Weingast (1997), in which democracy survives because citizens can coordinate to punish leaders who cross constitutional limits. If that coordination becomes harder, the limits weaken.
This lens connects two literatures that rarely speak to each other. One is AI safety. The other is the study of democratic backsliding, best known through Levitsky and Ziblatt's (2018) account of how democracies erode through legal channels rather than coups. Read together, the risk is not that a machine takes power. It is that a technology quietly raises the cost of the coordination that keeps elected leaders honest.
There is also a clear link to Shoshana Zuboff's (2019) analysis of surveillance capitalism. Zuboff showed how behavioral data became a commercial asset. Botvinick asks what happens when the same capacity is available to the state in moments of crisis.
Where the book is more demanding
Its long-horizon analysis is necessarily speculative. Phase II rests on assumptions about the pace of technological change that readers may not share. The deliberative proposal also inherits the open questions of the deliberative democracy literature, notably James Fishkin's (2009) work on deliberative polling: who designs the process, who sets the agenda, and how results translate into binding decisions.
Practical implications for leadership
For governments, parties and organizations, the book suggests three questions worth putting on the agenda now:
- Emergency protocols. Which decisions could be taken under emergency powers in an AI-related crisis, and what external checks apply to them?
- Data and surveillance limits. What data does the organization hold, who can access it, and under what legal standard?
- Labor transition narratives. If automation reshapes employment in a sector, who explains it to the public, and with what plan? An unexplained transition is fertile ground for authoritarian messaging.
Verdict
AI and Political Freedom is a necessary corrective to a debate that has been too focused on fake content and too little on institutions. Its contribution is analytical: it gives decision-makers a framework to anticipate where pressure will build, before a crisis forces the question. Recommended for anyone responsible for public institutions, campaigns or corporate governance.
References
Botvinick, M. (2026). AI and political freedom: The risk to democracy and how to respond. Princeton University Press. https://muse.jhu.edu/book/150140
Fishkin, J. S. (2009). When the people speak: Deliberative democracy and public consultation. Oxford University Press.
Levitsky, S., & Ziblatt, D. (2018). How democracies die. Crown.
Weingast, B. R. (1997). The political foundations of democracy and the rule of law. American Political Science Review, 91(2), 245–263. https://www.cambridge.org/core/journals/american-political-science-review/article/abs/political-foundations-of-democracy-and-the-rule-of-the-law/564FB92EE1808897FD0041E25289CF1F
Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. PublicAffairs.
