If Existential Risk Is What Matters, Direct Democracy Is Not Optional

Nick Bostrom has spent a career making one argument with unusual persistence, and I think it is right. Here is the argument, and then the consequence for how we govern ourselves that I have not seen him draw.

The Arithmetic of Extinction

Bostrom's claim is that even a small reduction in the probability of an existential catastrophe is worth far more than very large gains in almost anything else: curing diseases, extending lives, ending poverty.

The reasoning is not sentimental; it is arithmetic. A catastrophe that kills everyone does not merely kill the people alive. It forecloses every person who would ever have lived and everything they would have done. Weigh that against a cure for cancer, which saves an enormous number of lives but leaves the future intact, and the cure is a rounding error. Even if you discount the future heavily, even if you count only a few centuries, the numbers are lopsided enough that a fraction of a percent off the extinction risk beats nearly any other project.

You can quarrel with the framework. But I find that the people who quarrel with it have not usually done the sum; they have recoiled from doing it, which is a different thing, and one worth a post of its own. Grant the arithmetic and a great deal follows.

Mobilisation Is the Variable We Control

Bostrom's second point is that our capacity to mobilise as a species against a threat is itself one of the most important factors in the risk.

This is worth expanding, because it is where the argument stops being about probabilities and starts being about institutions. The existential risks we face are mostly not asteroid strikes, which do not care what we do. They are engineered pandemics, nuclear war, runaway climate, unaligned artificial intelligence: risks that are created by human action and that could be reduced by human action, if the action were coordinated and fast.

So the question is not only "how likely is the catastrophe" but "when it starts to become visible, how quickly can we move, and how hard?" A species that can see a threat, decide, and act within months is in a different risk category from one that sees the threat and spends a decade arguing about who will pay. The mobilisation capacity is a multiplier on everything else.

And unlike the asteroid, it is a variable we choose.

Where Representative Systems Fail This Test

Now look at the machinery we actually have for mobilising.

A representative legislature facing a slow, diffuse, expensive threat has every incentive to wait. The costs of action are immediate and concentrated on industries that fund campaigns; the benefits are deferred, spread across everyone, and consist of a disaster that then does not happen and so never appears in anyone's account. I have written about this pattern in Jared Diamond's terms: societies that survived their crises were the ones that could identify a danger and change course quickly, even against long custom, and the control of the few is exactly what slows that response.

Add the international layer and it gets worse. Existential threats are global; representative governments are national and answer to national interests in a game where defecting from a collective response pays. The result is treaties that are negotiated for years and honoured selectively.

Speed and force are the two things the situation demands, and they are the two things a system built on periodic elections, lobbying, and coalition arithmetic is worst at producing.

What Direct Decision Adds

Direct democracy does not make the public wiser about pandemics or artificial intelligence. That is not the claim.

The claim is about speed and force. A decision taken directly by the whole population has a legitimacy that no legislature can match, and legitimacy is what allows a society to act hard and fast without tearing itself apart. When the people who will bear the costs of a response have themselves voted for it, the response does not have to be negotiated with every interest group afterwards, does not have to survive the next election, and cannot be dismissed as the project of one party. It can simply be done.

It also removes the specific brake that stops legislatures: the concentrated interest with a stake in delay. There is no one to lobby when the decision is spread across millions of voters, most of whom have no career riding on the answer.

And direct decision at the level of a whole population is the only mechanism I know of that could scale up to the level the threats require. A world in which several large countries decided their big questions directly would be a world in which a global response could be assembled by asking the publics of those countries, rather than by waiting for their governments to find a moment when agreement was electorally safe.

The Conclusion

Put the pieces together. If existential risk dominates every other concern, then the capacity to mobilise against it is among the most valuable things a civilisation can have. If representative systems are structurally slow and weak in exactly the situations that matter, then the system of government is itself a large component of the risk. And if direct decision supplies the speed and legitimacy the situation needs, then moving toward it is not a political preference among others.

It is a safety measure, and it should be argued for the way we argue for any other. I have made the case for direct democracy on many grounds on this blog. This is the one I would put first if I had to keep only one.