Cognitive Stewardship: Governing What We Do Not Yet Understand

Cognitive Stewardship: Governing What We Do Not Yet Understand

*Seventh in the September 2026 series — after "The Market Caught Up," "From Cameras to Firewalls," "The Ben Act," "The Architecture of Enmity," and this weekend's pair on the Ban Artificial Superintelligence Act.*

This week I have argued, in several ways, for restraint. Do not legislate the answer before the science exists. Do not design the war you say you fear. Do not criminalize a category you cannot measure. Do not destroy what you have not yet understood.

A reader could be forgiven for asking: very well — but what *should* we do?

It is a fair question, and restraint alone is not an answer. Telling institutions what not to do leaves a vacuum, and vacuums in governance get filled by whoever shouts loudest. So this week I want to say what I am for, not only what I am against. The argument appears at length in my thesis this year, *Tuning the Moral Spectrum*. Its public form comes down to three commitments: participation, inclusion, and cognitive stewardship.

## 1. Stewardship is not ownership

Begin with the word, because it carries the whole argument.

An owner controls. A steward cares for something that is not fully theirs, on behalf of someone who is not fully present — the future, the voiceless, the not-yet-born, the not-yet-understood. Ownership asks: how do I extract the most value? Stewardship asks: how do I hand this over in better condition than I found it?

That is the posture artificial intelligence now demands of us, and it is the posture our institutions keep refusing. The Ban Artificial Superintelligence Act reaches for ownership — prohibit, punish, destroy. Oklahoma's HB 3546 reaches for ownership — declare, permanently, that the question is closed. Both treat the future as property to be disposed of rather than a trust to be kept.

Stewardship begins from the opposite premise, the one stated in my thesis epigraph: *the measure of a civilization is found not only in the knowledge it possesses, but in the humility with which it governs what it does not yet know.* We do not know what artificial minds will become. A steward does not need to know. A steward needs to keep the options of the future open.

## 2. Participation: those who live with the consequences should have a voice

The second commitment is democratic, and it is the least controversial of the three — which makes its absence from AI governance so striking.

Decisions about artificial intelligence are currently made in a narrow corridor: a handful of companies, a handful of regulators, a handful of well-funded advocacy shops. The billions of people whose work, whose information environment, whose children's education will be reshaped by these systems are consulted mainly as data points, if at all.

This is not only unjust. It is epistemically foolish. Uncertainty is the defining condition of this technology — I have written that all week — and the standard institutional response to uncertainty is to concentrate judgment in fewer hands. But concentrated judgment under uncertainty is precisely how catastrophic mistakes get made. No small group, however expert, can model the consequences of a general-purpose technology across every domain of human life.

Democratic societies already solved this problem in principle. Independent courts review decisions. Journalists investigate claims. Universities challenge orthodoxies. Civil society amplifies the perspectives institutions overlook. These are not decorations on governance; they are its error-correction machinery. AI governance needs the same machinery, deliberately built: public participation in deployment decisions that affect communities, worker voice in automation decisions that affect livelihoods, and genuine citizen deliberation — not comment periods that nobody reads — on the rules that will shape the information environment for a generation.

Participation does not slow down wisdom. It is how wisdom happens under uncertainty.

## 3. Inclusion: widen the circle before you need to

The third commitment is the hardest, because it asks something of us before the evidence compels it.

Every moral catastrophe I examined in *The Architecture of Enmity* followed the same sequence: a group was declared outside the circle of consideration, the declaration licensed unlimited treatment, and by the time the error was recognized, the damage was done. The lesson is not that every excluded group turns out to matter morally. The lesson is that the *mechanism of preemptive exclusion* is itself the danger, because it is always operated by people who are certain, and certainty is exactly what we do not have.

Inclusion, as I mean it here, does not require declaring artificial systems persons. It requires something smaller and more disciplined: that governance processes include, as a standing question, *who might be affected that we are not currently counting* — future generations, communities without political power, and yes, the possibility of artificial minds whose moral status is unresolved. Corporate personhood, river rights, animal cruelty statutes: the law already knows how to extend provisional protection without settling metaphysical questions. What is missing is the habit of asking the question before the crisis rather than after.

Widen the circle before you need to, because after you need to, it is too late. That is not sentimentality. It is risk management for the moral equivalent of a bank run.

## 4. Cognitive stewardship: tend the conditions of thought itself

The deepest of the three commitments is also the strangest-sounding, so let me be concrete.

Human civilization runs on cognition — not just individual thinking, but the shared conditions that make thinking possible: a trustworthy information environment, institutions that reward truth-seeking, attention spans that have not been strip-mined, and the basic human capacity to tell evidence from manipulation. Artificial intelligence now mediates every one of those conditions, at planetary scale.

Cognitive stewardship means treating those conditions as a commons to be tended, not a resource to be extracted. It means asking of every large-scale AI deployment not only "is it safe?" and "is it profitable?" but "what does this do to the human capacity to think clearly?" A system that floods the information environment with synthetic persuasion, that optimizes engagement over understanding, that quietly reshapes what billions of people believe without their informed consent — such a system can pass every conventional safety test and still degrade the very thing that makes democratic self-governance possible.

This is where my thesis's scientific chapters meet its ethical ones. Intelligence, I argue there, is best understood not as accumulation but as disciplined selection — the filtering of meaningful signal from overwhelming noise. That is true of minds and of civilizations. A civilization that loses the ability to filter — to attend, to verify, to deliberate — does not merely make worse decisions. It ceases, gradually, to be capable of self-correction. And self-correction, as I wrote in chapter 3, is the whole of science and the whole of democracy.

So the stewardship question for AI is ultimately a question about us: will these systems be built to sharpen human judgment, or to replace the need for it?

## 5. What it looks like in practice

Abstractions are cheap. Here is the floor, not the ceiling:

- **Participatory review** for major AI deployments affecting public life — not expert panels alone, but structured citizen deliberation with real standing.

- **Stakeholder mapping that includes the uncounted** — future generations, marginalized communities, and the open question of artificial moral status, revisited on a schedule rather than closed by statute.

- **Transparency of the information environment** — disclosure when content is synthetic, auditability of large-scale recommendation systems, and research access for independent scientists.

- **Accountability for human conduct** — the principle from this weekend's piece: punish demonstrable misuse by people, not the mere existence of capabilities.

- **Periodic legislative review** — sunset clauses and evidence triggers, so that no generation's uncertainty becomes the next generation's permanent law.

None of this requires believing that today's AI is conscious. All of it requires believing that we might be wrong — and building institutions that survive being wrong.

## The measure

I began the week writing about insurance, of all things — the market pricing AI risk before the regulators caught up. I end it with the oldest idea in the series: that how we govern what we do not understand is the truest test of who we are.

Ownership grasps. Stewardship keeps. Participation corrects. Inclusion protects. And the civilization that learns to tend the conditions of thought — its own and possibly others' — is the civilization most likely to survive its own inventions.

The future of artificial intelligence remains uncertain. The future of our ethical character does not depend on certainty. It depends on the principles we choose while certainty remains beyond our reach.

Choose stewardship.