The New Machinery of Rule

Every state is artificial. A constitution is not a forest, a river, or a season. It is a human construction: laws, courts, public offices, elections, rights, obligations, and institutions intended to allow strangers to govern a shared world without submitting entirely to force.

But another kind of artificial state is being built around us. It is not made of citizens, deliberation, or consent. It is made of data centers, corporate media conglomerates, surveillance systems, social-media feeds, predictive algorithms, AI agents, behavioral profiles, bots, and private terms of service. It does not ask people to govern themselves. It asks them to live under systems they neither own nor understand—and from which they cannot easily withdraw without losing access to work, communication, housing, credit, insurance, public information, or help.

In The Rise and Fall of the Artificial State, the historian Jill Lepore names this emerging order. Her argument is not that machines have suddenly become sovereign, or that a secret cabal directs every technological change. The Artificial State is more visible, more mundane, and more dangerous than that. It is the gradual replacement of democratic judgment with prediction, public deliberation with behavioral management, and citizenship with fine-grained forms of corporate control. Politics is converted into data. Public life becomes content. People become profiles.

Lepore’s Artificial State is a fantasy of escape from the burdens of human self-government. It imagines that difficult political questions can be made easier—or disappear altogether—when they are handed over to machines, markets, platforms, and private firms. But no algorithm abolishes politics. It hides political choices inside code. No predictive model removes power from human hands. It transfers power to those who own the model, choose the data, set the objectives, define the categories, and decide which errors are acceptable.

This matters under any conditions. It becomes existential in an age of biospheric collapse. By “biospheric collapse,” I mean not a single apocalyptic event but the compounding destabilization of climate, ecosystems, food systems, water supplies, public health, and infrastructure on which organized social life depends.

The coming decades will force collective decisions about water, food, electricity, housing, public health, insurance, evacuation, managed retreat, migration, debt, infrastructure, and ecosystems already under pressure. A society facing lethal heat, wildfire smoke, crop failures, flood damage, insurance withdrawal, and the growing instability of its life-support systems cannot afford to surrender its capacity for shared reality. Yet the emerging Artificial State profits when public attention is fragmented, when collective problems are individualized, and when democratic judgment is displaced by private systems of prediction and control.

Prediction Is Not Consent

Lepore begins with an insight that reaches back to Thomas Hobbes and Jean-Jacques Rousseau. States are artifices: human inventions built to organize collective life. Their legitimacy, at least in the democratic tradition, depends on the consent of the governed. People must be able to understand the institutions that rule them, challenge decisions, replace representatives, organize with one another, and appeal to shared principles of law and justice.

The Artificial State is something else. It is a political order in which private computational systems organize public behavior, public discourse, and public administration. Its infrastructure consists of targeted messaging, predictive algorithms, social-media campaigns, bots, AI agents, data collection, automated persuasion, and corporate-owned digital architecture. What had once been a citizen’s participation in public life becomes a measurable stream of clicks, reactions, purchases, location data, browsing behavior, and emotional responses.

The liberal-democratic state requires consent. It depends on people reasoning together, however imperfectly, about a common world. The Artificial State offers prediction instead: the promise that people can be known, sorted, influenced, and governed without being persuaded. It replaces the uncertainty of public judgment with the apparent certainty of calculation.

That promise is seductive. Democracy is slow. It requires argument, compromise, investigation, public meetings, elections, accountability, and the patience to treat people with different interests as political equals. Algorithms seem faster. Markets appear more efficient. A machine can process an application, rank a risk, flag a claim, route a call, target an advertisement, draft a message, identify a pattern, and predict a response at a scale no human institution can match.

But speed is not justice. Prediction is not consent. Efficiency is not legitimacy.

A public agency that decides who receives housing, health care, unemployment benefits, disaster aid, disability support, or relocation assistance is making political and moral choices. A private insurance model that determines who can afford coverage is making political and moral choices. A social-media platform that decides what people see, what they fear, what angers them, and which claims spread furthest is making political and moral choices. The fact that these decisions emerge from data systems rather than legislative chambers does not make them neutral. It makes them harder to see.

Automated decision systems are already used to assess eligibility for public benefits, detect alleged fraud, verify identity, and allocate public resources. When their data, rules, or error rates are opaque, people can lose access to assistance without a meaningful explanation or practical route to challenge the decision.

Consider a person denied food assistance, disability support, or disaster aid by an automated screening system. If the system cannot offer an intelligible reason, if its inputs cannot be inspected, and if the appeal process is inaccessible or too slow to correct an error, this is not merely a technical failure. It is public power exercised without a meaningful right of contest.

The danger is not that machines may someday become intelligent enough to dominate human beings. The more immediate danger is that people will accept a political order in which essential decisions are made by systems they cannot inspect, appeal, vote against, or meaningfully control.

From Viewers to Users

The Artificial State did not begin with artificial intelligence. It grew from older systems of corporate power over attention.

For decades, American public life has been shaped by media consolidation and private ownership of major communication systems, as well as the advertising-dependent economic pressures contributing to the decline of local news. Pew Research Center documents sharp drops in newspaper circulation and advertising revenue alongside the closure of thousands of newspapers. This does not mean every corporate journalist is dishonest, every newsroom operates as propaganda, or all media outlets are morally equivalent. It means the public sphere has been increasingly organized inside institutions whose survival depends on revenue, audience capture, corporate access, political influence, and the profitable management of attention.

Fox News is one important example. It did not invent propaganda, corporate ownership, or political manipulation. But it demonstrated how a commercial cable network could fuse news, entertainment, grievance, partisan identity, and political strategy into a durable emotional world. This is infotainment: political communication shaped as spectacle, personality, conflict, and commercial attention. Its product was a permanent sense of enlistment: viewers were invited to feel besieged, betrayed, morally superior, and surrounded by enemies.

The political effect of such a system is not limited to persuading viewers of any one claim. It changes the emotional form of politics. Compromise becomes surrender. Expertise becomes elitism. Public institutions become conspiracies. Opponents become enemies. A complex society becomes a melodrama in which every problem has a villain and every setback proves that someone is coming for your way of life.

Long before Silicon Valley, concentrated media power shaped national agendas and political identities. Lepore places today’s digital oligarchs in that longer history, noting the influence wielded by media barons and ideological broadcasters, including Roger Ailes and Rupert Murdoch. Corporate control of a consolidated, increasingly uncompetitive press is not a democratic achievement. It is a concentration of power over the stories through which people interpret their world.

Cable news could tell millions of people the same story every night. It could turn politics into spectacle and spectacle into identity. The platform economy changed both the scale and precision of that project: it transformed mass audiences into individually profiled users, able to be sorted, targeted, tested, tracked, and influenced in real time.

Corporate mass media taught people to become audiences. The platform economy taught corporations to turn audiences into data.

A cable network builds an audience. A platform fragments that audience into micro-publics—small, algorithmically segmented groups exposed to different messages, stories, emotional cues, and political frames—and continuously tests what will capture and hold each group’s attention.

Political microtargeting makes this fragmentation more consequential. Rather than address a public with one contestable argument, campaigns and platforms can use personal and behavioral data to tailor persuasive messages to narrowly defined groups, including messages calibrated to perceived psychological vulnerabilities.

A cable network can cultivate ideological loyalty. A platform can identify vulnerability, monitor reaction, amplify grievance, personalize fear, and target messages according to what produces the strongest engagement.

The older media system sold viewers to advertisers. The newer system sells predictions about behavior. It seeks not merely to know what people watch but to anticipate what they will do, believe, buy, fear, and keep watching.

Lepore’s Artificial State is built from this transformation. The public sphere is no longer merely influenced by corporations. It is increasingly engineered through privately owned systems that collect data from citizens, manipulate the conditions of attention, and reduce democratic participation to a sequence of measurable engagements.

The Degradation of Citizenship

The danger is not that Americans have suddenly become incapable of thought. It is that the conditions under which thought becomes politically effective have been systematically degraded.

A citizen requires more than an opinion. A citizen requires time, education, credible information, material security, and the ability to distinguish evidence from spectacle, disagreement from treason, and public life from partisan performance. A citizen also needs some confidence that institutions can be understood, challenged, repaired, and made to serve the public rather than merely the wealthy and well connected.

Those conditions have been weakened from several directions at once.

Economic insecurity narrows time and attention. A person struggling with rent, health care, debt, child care, unstable work, transportation, food prices, or disaster recovery does not possess the same capacity for sustained civic participation as someone insulated from those pressures. Local journalism has weakened, leaving people with less reliable information about school boards, public budgets, water districts, housing policy, hospital closures, utility regulation, and the everyday decisions that shape their lives. Civic education has deteriorated. The most recent national civics assessment found that only 22% of U.S. eighth graders reached the NAEP Proficient benchmark.

Meanwhile, the information environment has become louder, faster, more fragmented, and harder to verify. The Reuters Institute’s 2026 Digital News Report found that only 25% of Americans said they trusted most news most of the time. Distrust can be healthy when it leads people to investigate power. But generalized distrust—where no institution, source, fact, or authority can be believed—creates a vacuum in which emotional certainty, conspiracy, tribal loyalty, and charismatic manipulation flourish.

The Artificial State does not need every person to believe the same lie. It needs something more useful: people too distracted to follow power, too exhausted to participate, too economically insecure to take risks, too isolated to organize, too angry to deliberate, and too uncertain that truth can be known to demand a different future.

This is not simply ignorance. It is civic exhaustion.

A public constantly stimulated, frightened, humiliated, distracted, and economically strained becomes easier to govern through symbols than through arguments. Neighbors become enemies, social problems become moral panics, institutional failure becomes conspiracy, and democratic compromise becomes betrayal.

The citizen most useful to the Artificial State is not necessarily the one who believes every falsehood. It is the person who has lost confidence that truth can be known, public institutions can be repaired, collective action can matter, or anyone outside their political tribe deserves to be heard.

In that condition, politics no longer asks people to deliberate over a shared world. It asks them to choose which stream of manipulation feels most like home.

Neither Orwell Nor Huxley: The Dystopia We Chose

Orwell and Huxley imagined different nightmares of political control. Orwell imagined a state that ruled through surveillance, fear, censorship, official enemies, and the deliberate falsification of reality. Huxley imagined a society pacified through pleasure, distraction, consumption, entertainment, and the willing exchange of freedom for comfort.

The Artificial State resembles both, but it is identical to neither. It is not governed by a single Party, leader, or Ministry of Truth. Nor is it maintained chiefly by a population contentedly anesthetized into obedience. It is built through corporate-owned communication systems, partisan media, algorithmic prediction, targeted advertising, surveillance data, platform dependency, political microtargeting, and the mass production of digital noise.

It does not need to force every person to watch a telescreen. People carry the screen voluntarily. They pay for it, sleep beside it, consult it for news, work, intimacy, entertainment, navigation, medical advice, banking, and political identity. They supply it with their movements, preferences, fears, friendships, purchases, photographs, and private thoughts. The screen does not merely watch them. It learns what will keep them watching.

This is the dystopia we chose—not because most people knowingly voted for tyranny, but because convenience, connection, entertainment, speed, personalization, and apparent freedom were repeatedly allowed to substitute for privacy, public accountability, shared reality, and democratic control.

The danger is not merely that people will lose privacy, freedom, or the ability to distinguish truth from falsehood. These capacities are being lost as the biosphere becomes less stable: as heat grows deadlier, water less reliable, forests burn, food systems falter, coastlines retreat, and public institutions are asked to manage disasters they can no longer fully repair.

Orwell and Huxley imagined systems of control. The nightmare now emerging is control amid ecological breakdown: a population fragmented into algorithmic tribes, exhausted by spectacle, and unable to agree on the reality of its own conditions.

The deeper danger is the erosion of citizenship itself—the loss of the attention, judgment, solidarity, and democratic capacity required while the living world is pushed toward irreversible change.

The Climate of Manipulation

The consequences of this political degradation reach far beyond elections.

A society confronting biospheric destabilization must be able to recognize physical reality. It must understand that heat is not merely inconvenient weather when it overwhelms hospitals, threatens workers, strains electrical systems, and turns cooling into a survival necessity. It must understand that wildfire smoke is not simply a seasonal nuisance when it enters lungs, schools, homes, workplaces, and medical systems. It must understand that insurance withdrawal, water scarcity, crop stress, flood damage, collapsing infrastructure, and repeated disaster losses are not isolated personal failures. They are connected expressions of a climate and ecological system becoming less stable.

Climate-related health harms are already increasing in severity and geographic reach in the United States, while extreme heat, wildfire smoke, flooding, and other climate hazards compound risks to cardiovascular, respiratory, mental-health, and public-health systems. Smoke events require organized public-health communication, preparation, and protective measures; they are not problems households can reliably solve through individual consumer choices.

The coming decades will force decisions that are not merely about distributing hardship more fairly. They concern the material conditions under which organized civilization can continue to exist. The issue is no longer simply how to protect a familiar world from disruption. It is whether human societies can preserve the ecological, material, and political conditions required to remain organized and habitable at all.

• Which regions of the planet will remain safely habitable, and for whom, as extreme heat, water scarcity, fires, floods, sea-level rise, and ecosystem breakdown redraw the boundaries of ordinary life?
• How will billions of people be fed reliably when heat, drought, floods, soil degradation, pest outbreaks, and disrupted seasons destabilize harvests—and when crop losses strike several major food-producing regions at once?
• How will cities sustain water, sanitation, electricity, hospitals, housing, and public order when recurrent disasters damage the same systems faster than they can be repaired?
• Where will displaced people live when coasts, drylands, agricultural regions, and overheated cities become progressively harder to inhabit, insure, govern, or rebuild?
• What happens when the labor needed to grow food, construct housing, maintain grids, fight fires, and restore damaged infrastructure becomes dangerous or impossible in prolonged heat?
• Which ecosystems must be protected and restored because they are not amenities but living systems—watersheds, soils, forests, wetlands, fisheries, and pollinators—on which human survival depends?
• What forms of fossil extraction, luxury consumption, speculative development, water-intensive industry, and private accumulation must end because preserving them would destroy the conditions of life for everyone else?

These are questions of survival, continuity, and collective self-government. They cannot be resolved by a market optimized for short-term returns, nor by opaque systems that classify people as risks, costs, consumers, or data points. They demand democratic institutions capable of seeing a shared future, setting limits, distributing sacrifice and security fairly, and acting before collapse narrows every remaining choice.

In a destabilizing climate, these systems will not merely distribute services more efficiently. They will help determine which places remain viable, which communities can recover, and whose lives are treated as affordable losses. A risk model may judge a neighborhood too costly to defend after repeated floods, fires, heat emergencies, or water shortages. An insurer may withdraw coverage, making homes impossible to finance, sell, repair, or rebuild. A city may use cost-benefit models to decide where hospitals, transit, cooling networks, water systems, and flood protection will be maintained—and where decline will be managed instead.

This process is already underway. U.S. Treasury data show that homeowners-insurance nonrenewal rates are substantially higher in ZIP codes facing the greatest expected losses from climate-related hazards. When insurance disappears, a home can become harder to finance, sell, repair, or rebuild; a private model of risk begins to shape whether an entire community remains economically viable.

Each conclusion may arrive in the language of technical necessity. It remains a political judgment about whose safety, mobility, property, and future matter. Data will be essential for forecasting hazards, directing aid, protecting food and water systems, and maintaining infrastructure. But data cannot decide what a society owes people when their homes become progressively harder to inhabit. It cannot determine which losses are acceptable, whether a community has a claim to remain, or whether life must yield to property values, municipal balance sheets, and private return.

The Artificial State offers an answer of terrible convenience: replace public judgment with optimization. But optimization means choosing a goal and accepting what falls outside it. A system can optimize insurer profitability by withdrawing coverage, municipal solvency by abandoning infrastructure, or energy reliability for affluent districts by treating poorer communities as interruptible demand. It can optimize survival for a protected minority while converting everyone else into a managed population of risk.

No algorithm escapes politics. Every model defines what counts as value, what loss is acceptable, and whose future matters.

Who chooses the goal? Who decides whether the model values profit, property, public health, equal protection, ecological repair, democratic continuity, or human survival? Who determines the acceptable number of heat deaths, failed harvests, displaced families, uninsured regions, or sacrificed communities? Who can inspect the assumptions, appeal the decision, or reject the premise that their life is an efficient loss?

When such questions disappear into proprietary systems, politics does not vanish. It becomes invisible just as its consequences become final. Citizens become subjects of systems that may determine whether they can remain in their homes, feed their families, receive care, move safely, or inhabit a functioning community.

Attention-driven platforms can profit from outrage, fear, and distraction. Fossil interests profit from delay. Political opportunists profit from scapegoating. The Artificial State gives all three a common infrastructure: it turns material breakdown into market signals, fear into engagement, abandonment into a risk score, and collective failure into individualized loss.

From Citizens to Subjects

Jill Lepore does not argue that every computer, database, algorithm, or artificial-intelligence system is a threat to humanity. Medicine, public health, scientific research, weather forecasting, emergency communication, environmental monitoring, education, agriculture, and urban planning can all benefit from computational tools. A society entering ecological crisis will need rigorous science, reliable data, resilient communication, and technologies capable of reducing harm.

The machinery is not the danger by itself. The danger is allowing its owners, incentives, and classifications to become the hidden constitution of public life: the unvoted rules that decide whose claims are legible, whose needs are urgent, whose losses are tolerable, and whose future can be written off.

Democratic use of technology requires more than good intentions. It requires public rules, understandable reasons, independent audits, appeal rights, data minimization, meaningful human review, worker and community participation, and institutions able to overrule private systems when public survival is at stake.

Tools that help firefighters anticipate dangerous conditions, monitor water quality, identify grid failures, forecast heat risk, improve medical imaging, coordinate evacuation, or map flood exposure can protect people from harm. But they cannot assume democratic responsibility.

A model can identify a floodplain. It cannot decide whether residents are owed dignified relocation, secure housing, and a viable future elsewhere—or only abandonment after the water rises. A system can calculate heat risk. It cannot decide whether cooling is a luxury for those able to pay or a condition of survival. A platform can distribute emergency alerts. It cannot supply the housing, health care, water, transportation, income, and recovery systems that make an alert more than an announcement of danger. An algorithm can estimate an insurance loss. It cannot decide whether a family should be trapped in an uninsurable home, whether a renter should be displaced after disaster, or whether an entire town should be allowed to lose its tax base, schools, clinics, and public services.

A machine can recognize a pattern. It cannot answer for what is done to the people caught inside it.

This is the political temptation of the Artificial State: to translate decisions about obligation, rights, survival, and abandonment into administrative outputs. A family becomes a risk profile. A neighborhood becomes a cost center. A flooded town becomes a negative return on investment. A person fleeing heat, hunger, fire, or rising water becomes a burden to be routed, surveilled, detained, priced, or excluded. The language of efficiency makes these outcomes sound inevitable. It turns choices into scores, dispossession into procedure, and political abandonment into a technical result.

The resulting dystopia will not necessarily announce itself with a dictator, a uniform, or a single command center. It may arrive as an insurance cancellation that makes a home unsellable; an unaffordable utility bill during a heat emergency; an automated denial of aid; a closed clinic; a failing transformer; a contaminated well; an evacuation order with nowhere safe to go; or a platform that keeps people furious at one another while the institutions capable of protecting them are sold, hollowed out, or permitted to fail. It will feel less like conquest than attrition: a narrowing of options, year after year, until entire populations are told—by price, risk score, service cutoff, or absence—that they no longer fit inside the world being maintained.

A functioning democracy is not a feed. A citizen is not a user. A population is not a market segment. And a living planet is not a warehouse for data centers, a supply of free inputs, or a backdrop for corporations that imagine themselves exempt from the obligations of Earth.

The Artificial State promises efficiency without politics, administration without discretion, communication without community, and intelligence without responsibility. In practice, it offers a society in which people are endlessly measured but increasingly unrepresented; constantly addressed but never heard; predicted in exquisite detail but deprived of meaningful power over the systems that shape the terms of their survival.

No society can plan for a future it has been trained not to see. A people unable to distinguish evidence from spectacle, material danger from manufactured panic, or reality from the stories designed to monetize and manage them cannot deliberate about survival. It can only react—late, divided, and vulnerable—to crises it no longer possesses the common language to name.

Machines can make government faster. They can sort claims, score risks, rank priorities, automate messages, and enforce decisions at a scale no human bureaucracy could match. But speed does not produce judgment. Prediction does not produce consent. Engagement does not produce citizenship. Corporate infrastructure does not become public authority simply because people cannot live without it.

In an age when biospheric collapse is no longer an abstraction but a condition unfolding unevenly across the planet, those distinctions become matters of survival. Either people retain institutions capable of deliberation, solidarity, limits, democratic control, and repair—or they become subjects of systems designed not to prevent collapse, but to sort, price, contain, and administer its human casualties, one individualized loss at a time.