Ephraim Agbo's Blog: When Institutions Falter: Francis Fukuyama on Democracy, Power and the New Age of Machines

By Ephraim Agbo

Recently I listened to an interview with Professor Francis Fukuyama — the political economist whose The End of History and the Last Man shaped the way many of us understood the post–Cold War world. I read that book as a young student and it lodged in my thinking. Hearing Fukuyama speak decades later felt like meeting an old teacher who had kept thinking about the problem long after the lecture ended. What follows is an extended, intensely personal unpacking of what he said, why it matters, where I think the analysis is strongest, where I worry about gaps, and what practical steps flow from the diagnosis.

“We’re seeing that erosion taking place before our eyes.” — Francis Fukuyama

That line framed the whole interview for me. It’s blunt, simple, and terrifying because erosion rarely arrives as spectacle; it arrives as attrition. Below I trace the layers of that attrition — constitutional, informational, economic, and technological — and then work through what defense might look like in practice.

1) How democracies quietly lose their protections

Fukuyama’s central political claim is not that democracies spontaneously collapse, but that they weaken: institutions designed to constrain power are slowly hollowed out. He pointed to Viktor Orbán’s Hungary as a useful comparator — not because Hungary is identical to the United States, but because Orbán’s playbook shows how legal and administrative mechanisms can be repurposed over a decade to concentrate authority and neutralize opposition.

Why does this matter? Two reasons:

Normalizing exceptions: Democracies survive on norms as much as laws. When governments normalise “small” exceptions — emergency powers, aggressive immigration enforcement without due process, executive fiat — they lower the cost of future violations. What starts as a “temporary” measure begets another, and then another.
Delegitimizing constraints: Once a political movement treats courts, civil-service protections, or independent agencies as partisan obstacles, it becomes politically acceptable to attack them. The rhetorical delegitimization precedes legal capture.

Listening to Fukuyama, I kept thinking of concrete examples he mentioned: masked agents snatching people off the streets, detentions with little due process, and the use of law enforcement as a theater of deterrence rather than justice. These are not theoretical constructs. They are ways democratic norms get bent until they break.

The presidency, the “unitary executive” and the new constitutional logic

A long stretch of the interview was devoted to the use—and abuse—of executive power. Fukuyama draws an important distinction between emergency powers that are ultimately checked by courts (FDR is his historical example) and a modern posture that openly attacks judicial review as illegitimate.

Two linked ideas matter here:

Unitary-executive theory: This legal theory—revived and popular in some conservative circles—argues for a strong, centralized executive with wide latitude over administrative decisions. Fukuyama thinks this is not academic hair-splitting; it is the doctrinal backbone for concentrating political control.
The performative presidency: Beyond law, the modern president can use rhetoric and the instruments of state to delegitimize opponents (courts, press, bureaucracies). Acting this way changes the incentives of public servants and can erode institutional independence even without formal legal changes.

The political consequence is clear: if you make it normal to treat judges as partisan actors, you reduce the ability of the judiciary to act as an effective check. If courts are treated as enemies, the risk of packing, bypassing, or simply ignoring courts becomes higher.

Strange affinities: why unlikely leaders can be politically similar

Fukuyama’s provocative point—“It’s appropriate that Putin and Trump meet”—grabbed my attention. He was not collapsing ideologies; he meant something more structural. Authoritarians (or would-be authoritarians) share a habit: instrumentalizing institutions for political ends. Whether through formal legal reform (as Orbán did) or through informal pressure and spectacle (as in other settings), the effect is similar: institutions are weakened and power is concentrated.

This is a subtle but crucial idea. Authoritarian mimicry isn’t about copying laws; it’s about copying habits of power — the inclination to treat institutions as tools rather than constraints. That mimicry can cross cultural and ideological divides and is therefore especially dangerous.

Campus politics, the conservative backlash, and the weaponization of “order”

Universities emerged as a microcosm in the interview. Fukuyama acknowledged left-wing excesses—protests and rhetoric that sometimes crossed lines—but he was more concerned with how the state uses disorder as a pretext to impose controls: regulating curricula, instructing universities on whom to admit or how to teach, and pursuing policies that erode academic autonomy.

Two lessons for me stood out:

Rights need boring work: Moral clarity on campus protests must be matched with organizational work: voter registration, student representation, legal defense funds. Activism without institution-building is vulnerable.
The state’s logic is strategic: When a government sees campus disorder as politically useful, it will escalate control. Protecting academic freedom therefore requires political alliances beyond campus — with courts, legislatures, and civil society.

The “loss of control” problem in AI and the collapse of epistemic trust

When Fukuyama described his Instagram feed — “more than half the videos I see are deepfakes” — the remark moved from anecdote to existential symptom. He described two separate but related threats from AI:

Epistemic erosion: Deepfakes, AI-generated audio and video, and other synthetic content undermine shared facts. Democracy relies on a common informational baseline; without it, public deliberation becomes impossible. Misinformation used to be noisy; now it can be convincingly false at scale.
Autonomous decision-making: Machines are increasingly delegated authority (drones, targeting algorithms, automated policing tools). When decisions that carry moral weight are delegated to machines, accountability and judgement are obscured. This is what Fukuyama calls the “loss of control” problem.

Combine these with geopolitical competition (a rush to deploy because of fear of falling behind a rival), and you have a dangerous incentive structure: proceed fast, regulate slowly, and accept collateral harm as the price of strategic advantage.

From a political standpoint, that is catastrophic. A state that cannot control informational integrity or the responsible use of force is a state in which democratic oversight is hollowed out.

Economic concentration, labor disruption and political fallout

Fukuyama linked AI to labor market disruptions that are already visible — entry-level roles replaced by automation, early career displacement. The socioeconomic consequence is not just unemployment; it is concentration of wealth and power in firms that own data and models.

The political consequence follows logically: economic power begets political leverage. If a handful of corporations control the platforms for information, employment, and essential infrastructure, then democratic choices become constrained by economic realities. Inequality breeds resentment, and resentment breeds appetite for shortcuts that promise immediate nationalistic or authoritarian fixes.

Machines and moral personhood: an important philosophical boundary

Fukuyama was skeptical about granting rights to machines. His key observation is simple: rights rest on experiences—consciousness, pain, fear, desire. Machines may simulate those states but simulation is not necessarily equivalent to experience. The moral claim that grounds rights (subjective well-being, vulnerability, interest) remains uniquely human for now.

This does not make machine policy any easier. The core problem isn’t whether a machine “deserves” rights; it’s how machines amplify existing human inequalities and distort the information and economic architectures on which democracies depend.

Where Fukuyama’s diagnosis is strongest — and where I want more

Strong points

The focus on gradual institutional erosion is crucial; it reframes our sense of danger and the remedies required (slow, sustained civic work rather than panic).
The coupling of political norms with legal structures is insightful: law without norm becomes brittle; norm without law is precarious.
The emphasis on epistemic risk from AI is timely and underappreciated in many political discussions.

Open questions / places I want more

Policy specificity. Fukuyama diagnoses brilliantly but is less prescriptive about concrete legislative architectures for AI governance, antitrust in the platform era, or judicial protections that can survive political polarization. Diagnosis needs to be matched with legislative design.
Comparative resilience. Why have some democracies weathered these pressures better than others? Fukuyama gestures to Asia’s democracies (Japan, South Korea, Taiwan) but we need deeper comparative analysis of institutional design features that confer resilience.
Economic safety nets. He rightly warns about inequality but stops short of describing the fiscal and labor policies (universal basic income pilots, job guarantees, retraining at scale) that could blunt political backlash.

Practical prescriptions — what governments, institutions and citizens should do

Fukuyama’s central moral was mobilizational: don’t flee, organize. I translate that into specific actions across four levels.

For policymakers and legislators

Enact AI governance frameworks that combine transparency, auditing, and liability for harms. Create independent AI oversight bodies with teeth.
Modernize antitrust to reflect data/network effects and platform dominance. Break up or regulate gatekeeper behavior that controls information and labor platforms.
Protect judicial independence with institutional safeguards (secure tenure models, public funding for court administration, norms that penalize political attacks on judges).
Fund public-interest media and verification labs to counter the erosion of trusted information sources.

For universities and educators

Embed media literacy across curricula — not optional modules but core competencies across disciplines.
Build institutional firewalls: legal resources, independent ombuds, and alliances with civil liberties organizations to defend academic freedom.
Translate protest into civic capacity: turn campus activism into voter registration drives, community organizing and durable coalitions.

For tech companies and engineers

Adopt robust provenance standards for generated media (cryptographic provenance, watermarking).
Prioritize safety over first-mover advantage: slow deployment of risky capabilities and open cooperation on standards, especially for military and surveillance tech.
Design for accountability: logging, explainability, human-in-the-loop systems for high-stakes decisions.

For citizens and civil society

Vote and organize locally. Local elections and school boards matter for institutional control.
Support independent journalism. Subscribe, donate, and volunteer for verification projects.
Practice media hygiene. Verify before sharing, diversify information sources, and support media-literacy programs in your community.

Scenarios and guardrails (a short strategic imagination)

Thinking strategically helps. Three plausible near-term scenarios illustrate different political futures:

Hardening of institutions: Strong civic mobilization, regulated tech deployment, and anti-concentration policies slow erosions. Democracy adapts and reasserts norms — an incremental but durable renewal.
Entrenchment of executive power + technological dominance: Executive aggrandizement continues, major platforms dominate information and labor, and geopolitical races justify rapid deployment — an era of managed democracies with limited pluralism.
Authoritarian contagion: Institutional erosion accelerates, courts are neutralized, disinformation creates permanent epistemic fractures — a global environment where democratic governance is episodic and fragile.

Our actions—policy, civic organizing, international coordination—determine which path becomes dominant. The point is not to forecast precisely but to orient strategy toward resilience.

Final reflections — why this moment demands patient courage

Listening to Fukuyama felt like a call to steady, unglamorous work. The threats are real and manifold — procedural erosion, epistemic collapse, economic concentration, technological acceleration — but none are magic. They are human problems born of human institutions and choices. That is the hope in his diagnosis: if humans made these systems, human imagination and political will can fix or at least slow the damage.

What stays with me most is the normative demand of democracy: it requires daily maintenance. It thrives not on heroic acts alone but on the boring, stubborn labor of institutions, courts, civic organizations, and an informed populace. Fugitive passions matter; durable infrastructures matter more.

If you take one thing from this reflection, let it be this: defend institutions not because they are antiquated relics but because they are the scaffolding that makes pluralistic politics possible. Protect the conditions for shared truth. Insist on economic rules that distribute opportunity. And above all, organize — patiently, persistently, and with moral clarity.

Ephraim Agbo's Blog

September 13, 2025

When Institutions Falter: Francis Fukuyama on Democracy, Power and the New Age of Machines