Earlier this month, San Francisco’s Hamilton Society held a debate on a question that would have sounded absurd not long ago: Does America need a Caesar?

Older Americans may scoff. Younger politically engaged Americans are less likely to. Many came of age after 9/11 amid corruption, institutional drift, and visible decline. Even a presidency that has slowed the Biden-era collapse has not repaired the deeper sense that the system no longer works.

America needs a path out of cybernetic collapse that does not end in machine rule or technologically enhanced autocracy.

From that premise, the temptation follows: If everything is broken, perhaps history’s oldest emergency remedy deserves a second look.

But one does not simply “have a Caesar.” Critics who dismiss the whole discussion as live-action role-playing or strongman wish-casting have a point. A generation trained to think and talk rather than act will naturally produce elaborate theories of action it has never practiced.

The Hamilton Society debate captured that tension. The motion that America needs a Caesar failed, 56% to 44%, though the pro-Caesar side gained five points after the arguments. I had agreed to argue against the motion before a scheduling conflict kept me away.

Nervous breakdown

That case is easy enough to make today. America does not need a Caesar right now. The disturbing question is whether conditions could change until it does.

The danger lies in those two little words: right now.

Right now, America is entering a period of radical uncertainty, driven in part by artificial intelligence. My feed is filling with nervous remarks from people working at the edge of the AI industry. Their concern confirms two things: Our near-term future remains unwritten, and our capacity to make plans is beginning to break down.

The threat is not only a race toward recursively self-improving intelligence. It is also the recursive destruction of our ability to understand what is happening while our political order comes apart.

RELATED: The path to America-first AI

yunping liang/iStock/Getty Images

The federal government faces a brutal dilemma. Its legitimacy depends partly on maintaining American strategic dominance. As AI advances, can Washington preserve that dominance without claiming the most powerful systems for itself and using them at its discretion?

Maintaining such an edge would not mean controlling only today’s strongest models. It would mean controlling tomorrow’s, then the next day’s, forever. The ultimate AI weapon is not one machine. It is the power to weaponize anything at any scale, from the planetary to the personal.

The technocratic answer is more cybernetics: tighter governance through self-improving systems of monitoring, prediction, and control. But we should be preparing for cybernetics itself to fail under the pressure of the moment.

Diminishing options

When a system can no longer govern itself, societies reach for familiar exits.

One is withdrawal: Head for the hills, build resilient communities, and live beyond the failing machinery. Some people should do exactly that. But if escape becomes a mass impulse, it stops being resilience and becomes flight. Poorly prepared internal refugees would remain trapped inside the same collapsing country.

A second option is spiritual authority. In periods of corruption, societies often turn to their holiest figures. America, however, is ill-equipped for that solution. Our churches are divided, many religious leaders suffer their own crisis of trust, and our constitutional order does not provide for theocracy.

That leaves the third option: one man entrusted with enough power to impose order on the people and, now, their machines. A ruler believed strong enough to bear the burden of writing a new law for a broken society.

In other words: Caesar.

What would have to happen before Americans not only wanted such a figure but believed they needed him? A broad majority would have to conclude that without one, the country would collapse and Americans would cease to be recognizably American.

RELATED: The question of the AI age: How much humanism is too much?

Moor Studio/iStock/Getty Images

The failure of the American experiment would leave us unable to agree on what regime should replace it. We might not even know what sort of people we had become. Under those conditions, an all-powerful AI could offer to decide for us.

If the choice appeared to be submission to the machine or submission to a man who promised to preserve both our humanity and our national identity, Caesar would suddenly look pretty attractive.

A civilization, if we can keep it

Here is the catch: Every historical Caesar has been pagan or Christian, and Christian Caesars ruled empires. America has imperial features, but it is not the kind of civilization likely to produce a Christian emperor. The Christian model for the ruler of one nation is the king, not the Caesar who governs many peoples.

That leaves the more plausible American outcome: a pagan Caesar.

Few Americans worship the classical gods. Politically, however, pagan gods were the gods of the regime — powers embodied by the city, the state, and ultimately the ruler. Caesar did not merely govern beneath them. He could become one of them.

Artificial intelligence now offers a lone ruler something no ancient emperor possessed: the prospect of near-divine control over information, movement, behavior, and even human cognition. AI could restore the oldest pagan political dream — the “gods of the city” made operational.

That is the point of departure for “Pagan.ai,” my essay in the new issue of Frontier.

America needs a path out of cybernetic collapse that does not end in civil anarchy or cyber theocracy. If we don’t like the Caesar we may someday believe we need, we should begin now to repair the conditions that would make him possible.

The best way to avoid Caesar is not to mock the question. It is to build a country that never has to ask it.