VII. A Civic Practice for the AI Age

What the evidence asks of us

The school trust ran for two hundred and forty years before its enforcement constituency arrived in serious form, and the asks for that two-and-a-half-century delay were many. Each generation that did not show up had its reasons, sometimes good ones and sometimes only the reasons of distraction and exhaustion that are the ordinary lot of any working adult, and the reasons compounded, generation upon generation, into the drift the rest of this paper has documented at length.

The aspirational structures the AI moment is now constructing will not have two and a half centuries to acquire their constituencies. They have, on the political timescale most of them are operating in, perhaps five to ten years before their architectures harden into the form they will keep. The white paper’s argument is not that this is a tragedy. It is that the historical evidence is now in, and the choices the next several years will produce can be informed by it — by the same evidence, drawn from the same 240-year natural experiment, applied with discipline to a set of structures whose founders are still in the room and still have time to listen.

What follows is what each of the white paper’s four primary readerships can do with that evidence.

For policymakers

Codify the seven pillars in any new aspirational structure before the structure is funded.

The federal architecture for the school trust was built across 125 years of admission acts. By 1910 it had reached the strongest formulation any American statute has produced — express trust language, breach-of-trust nullification, mandatory auction, separate funds per object, U.S. Attorney General enforcement.¹ The architecture works where it has been installed. What the historical record also shows, however, is that installing the architecture in arrears, after a structure has been operating for decades and after constituencies have grown around the diversions the structure permits, is harder by an order of magnitude than installing the architecture at the founding.

The operational rule for the policymaker is therefore narrow and clear. Every new sovereign wealth fund charter, every new climate trust enabling statute, every new compute-reserve authority, every AI Targeting Authority’s foundational document, should contain — at minimum — the seven elements named in Section V.² Codified fiduciary duties to the named beneficiary class. A restoration mechanism for diverted assets. Personal accountability for trustees. Direct distribution that produces visibility at the beneficiary level. An independent advocacy office statutorily separate from the management entity. Specialized legal defense available to the trust when its interests conflict with other state actors. Mandatory annual fiduciary education for trustees, year in and year out for as long as the structure exists.

None of these is novel. Each is in operation somewhere. The federal architecture is on the books in the 1910 New Mexico-Arizona Enabling Act; the operational architecture is on the books in Utah’s Title 53C and 53D.³ Replication is a matter of legislative drafting, not of design. The policymaker who declines to install the architecture, in 2027 or 2030 or 2035, will be making a choice the historical record will recognize for what it is — and the policymaker who does install it will be making a choice the beneficiaries fifty and a hundred years out will not be in a position to thank them for, but the historical record will.

For investors and philanthropists

Fund the rails, not only the application layer. Own the primitives.

The AI economy’s first decade has been dominated, in foundation strategy and venture capital alike, by the application layer. The aspirational-structure layer has received attention but limited capital, and the imbalance has consequences the school-trust evidence makes legible. The historical evidence is that the layer that matters most, over generational time, is the one that sets the architecture — and that the architecture, once set, is harder to redirect than its founders typically imagine. The framers of 1785 did not imagine, as they sat down to write the Land Ordinance, that two hundred and forty years on the most consequential debate in American education funding would still be turning on whether Section 16 of every township was a “grant” or a “trust.” But it was, and it is.

For foundation program officers and philanthropic strategists, the question is which structures, in 2026 and 2027, are entering their charter window. The chartering decisions produce architectural facts that will be operational in 2050, in 2075, and in 2125. Capital deployed at the chartering moment — in research support, in advocacy infrastructure, in comparative-policy analysis, in statutory drafting expertise — produces leverage that no later capital deployment can match. The school-trust history is, in this sense, an investment thesis. The structures with strong founding architectures perform; the structures with weak founding architectures consume capital across decades on remediation that does not fully recover the lost ground, the way Oregon has been consuming reform capital on the Common School Fund for the better part of fifty years without restoring the corpus to its design.

For investors, the analogous move is to recognize that the institutional infrastructure of the AI moment — the targeting authorities, the climate trusts, the data trusts, the compute reserves — is itself an asset class with generational duration. Allocations to that infrastructure, made at the chartering moment, are not philanthropy; they are the investment of capital into the structures that will determine the conditions under which all other AI-economy investment will operate over the next century. The investor who treats the architecture layer as the policy people’s problem, and not the investor’s problem, is making a thirty-year miscalculation that will compound at the same rate the underlying assets compound.

For citizens

Become part of the standing constituency.

The eighth element of the architecture — the one Section V identified as qualitatively different from the seven pillars — is the active enforcement community across generational time.⁴ The school trust’s failure modes were not, in their final form, failures of architecture. They were failures of the constituency around the architecture to keep showing up, generation after generation, as the founding generation tired and the second generation forgot why the founding generation had cared and the third generation never quite learned the story in the first place.

The 2026 environment changes this. AI-augmented research has broken the institutional monopoly on historical memory, giving ordinary citizens the analytic leverage to monitor complex, multi-generational assets in real time. A retired tree farmer in Coos County can, in 2026, operate at a level of analytic competence on school-trust matters that would have required a state archivist’s resources in 1986. The same is true for every other domain a standing constituency would attend to, which is the hopeful side of the same cognitive-technology shift Section VI describes.

The operational task for the citizen is to become part of the standing community. The institutional locus exists: America’s School Trust Library at schooltrusts.net. The Library is the first public-facing instance of the watchful crew this paper has named — the evidentiary archive, the weekly Newsroom, the per-state transparency reporting, the recruitment infrastructure for the deputy auditors and state correspondents who will, over the next decade, become the standing constituency the framers of 1785 could not engineer for.

A note on the register in which this work is being done.

Americans are slow to anger, devastating once roused, and conditioned by their country’s own history to expect that a great civic emergency will arrive announced. Pearl Harbor did what no amount of pre-war argument could do. December 7, 1941, is the canonical American example of a crisis whose arrival reset the political will of an inattentive country in a single afternoon, and the instinct to ask whether the school-trust failure, and the AI-era trust failures that the school-trust record predicts, will likewise be resolved by some galvanizing event is a natural one. It is also, on the historical evidence, the wrong instinct.

The school-trust record says the instinct is not just wrong; it is exactly inverted.

Drift-mode failures do not produce crisis-mode wake-up calls. Mississippi sold its sixteenth sections for a dollar an acre in the 1820s and there were no marches in the streets. California disposed of its school sections in 1853 and there was no national crisis, no editorial outrage in any paper that lasted more than a news cycle. Oregon let the Irreducible Fund fall to roughly twenty percent of its design across sixty years of quiet legislative friction, and Governor Chamberlain’s 1907 message — the document that gives us the “five or six times as large” number — produced no policy response, no legislative session called to address it, no governor in the next forty years willing to put it back on the agenda. The drift mode is the failure mode that does not generate a Pearl Harbor. That is the lesson the school-trust record offers; it is not an exception to it.

Which means: an AI-era perpetual climate trust that quietly compounds at four percent below design every year, an AI Targeting Authority whose mandate is gradually redefined by whichever administration’s appointees sit on its board this decade, a longevity escrow whose beneficiary class is rewritten under fiscal pressure in 2055 — none of these will arrive as crisis. The crisis will not arrive as crisis. The crisis is already arriving, in the quiet form drift always takes, while the country looks elsewhere.

The crisis won't arrive as crisis. The crisis is already arriving.

The right reference for the work this paper is part of is therefore not Franklin Roosevelt. It is Rachel Carson. Silent Spring in 1962 took a diffuse, structural, invisible harm — DDT working silently through ecosystems for two decades — and compressed it into a single legible argument that produced the Environmental Protection Agency in 1970.⁵ Carson did not wait for a crisis. She wrote the book that was the crisis, in calm scholarly voice, page by careful page, and the public found it unignorable. Eight years from book to agency. That is the model this paper takes for itself.

Eight years from book to agency. That is the model.

Translated to the work this paper sits inside: the Library is the crisis-legibility instrument. Its job — and the job of the constituency assembling around it — is not to ring an alarm bell at a sleeping country. It is to make a slow disaster fast enough to read. That is a different posture from Roosevelt’s wartime mobilization, and it is a different posture from polite scholarship, and it is the posture this paper has been written in throughout. It is calm voice telling the reader something genuinely alarming. The standing constituency this paper has been arguing for is what gives that voice the institutional durability to keep speaking until the message lands, in the way Carson’s voice kept speaking through the decade between her book and the agency that her book, in the end, produced.

The Library is the crisis-legibility instrument. Its job is to make a slow disaster fast enough to read.

Around the Library are the advocacy organizations through which the citizen’s work translates into policy, litigation, and education pressure: the Advocates for School Trust Lands (ASTL) at the national level; the Oregon Advocates for School Trust Lands (OASTL) and parallel state organizations; the developing pro bono legal community whose chartering this paper has been part of.⁶

For aspirational structures still in their charter windows — climate trusts, AI Targeting Authorities, the rest of the lengthening list — the analogous institutional locus does not yet exist. It will benefit, in its forming year, from the lessons the Library this paper hosts has been working out in real time: that the standing community does not assemble itself; that it requires a public-facing institutional anchor that holds the substrate; that the founders of the structure must build the institution before they can hope for the constituency, not the other way around.

What the standing constituency requires from each of its participants is not heroism. It is presence. The framers, at the end of Section II.A, knew they were building for a beneficiary class without political voice. They had no instrument with which to compensate for the dependency they were creating. Each generation of citizens since then has had the option of supplying that instrument by being there, and most generations have not. The school trust’s drift is, in part, the cumulative cost of generations that mostly did not.

The 2026 moment is when that pattern can change.

For the founders of the coming trusts

Write the seven pillars into the founding compact. Do not assume standing virtue. Engineer for its dissipation.

This is the audience the white paper has worked hardest to reach. The architects of the AI Targeting Authority, the climate trust, the longevity escrow, the data trust, the compute reserve — most of them not yet named, most of their structures not yet chartered — are the population in the position to make installation the default rather than the exception, on the schedule that has been set for them by the pace at which the AI-era trusts are being designed.

The case is the simple one. The seven pillars are derived from a 240-year natural experiment with several dozen test conditions.⁷ Their installation is empirically associated with the structures that hold their ground; their absence is empirically associated with the structures that drift. The standing-constituency requirement is empirically associated with the recoveries even within the failure cases, the way Utah’s recovery in the 1990s was empirically associated with the assembly of its enforcement community across the preceding decade. The architecture is the working manual the historical record has produced.

The founders of the coming trusts must avoid the single greatest oversight of 1785: assuming the moral clarity of the founding generation will automatically propagate to the next. It does not. It dissipates, on the historical evidence, within two or three generations at most, and sometimes within one. The structures that survived dissipation are those whose founders engineered for it: who made the architecture statutory rather than cultural, who provided enforcement mechanisms that did not depend on any single official’s continued attention, and who supplied the rails for a constituency to assemble around the structure across the time the structure would actually have to last.

The founders of the coming trusts have, in 2026, the historical record the eighteenth-century framers did not have. The choice they make with it will compound for centuries.

The choice

The argument is the simple one. We have a 250-year case study in how aspirational structures behave under intergenerational pressure. The case study tells us that drift is structural; that the structural defenses are knowable and installable; and that the enforcement constituency is qualitatively different from the architecture and is what makes the architecture work over the time horizons the architecture was meant to operate on. We are now, in 2026, building the most consequential aspirational structures in human history — structures whose beneficiaries are the future humans whose welfare we say we are designing for.

The choice is whether we absorb the lesson the school trust has been teaching, in installments, since 1785. It is whether we install the seven pillars at the chartering moment of each new structure, while installation is cheap. It is whether we make the standing constituency a precondition for the structure rather than a post-hoc remedy. It is whether the framers of 2026 take seriously what the framers of 1785 could not — the obligation to engineer for the dissipation of the virtue that animates the structure’s founding.

A forever gift to forever schools for a forever democracy was the eighteenth century’s working title for what the school trust intended to be.⁸ The phrase has held up. The structure beneath it has held up partially, in some places more than others, with a recovery community arriving late and assembling around what is left of what was set down. What the white paper’s argument now asks is that the analogous structures of the twenty-first century not require the same long delay. The instrument that the framers could not engineer for is, for the first time, within reach. The choice to use it is in front of us.

The school trust has carried the case as far as one institution can. The next institutions, on the historical evidence, will carry it farther — or they will not.

That choice is what this paper exists to put on the page.

—

Footnotes / source notes