Libertarianism

Morally Binding: The Super Intelligence Accord and the Problem of Self-Certified Power

Posted on by jdd

By JD DuRie, in collaboration with AI Matters

On September 29, President Trump invited six of the most powerful people in technology to lunch. Afterward he took them outside to sign a document. Present were Sundar Pichai of Google, Dario Amodei of Anthropic, Mark Zuckerberg of Meta, Greg Brockman of OpenAI, Elon Musk of xAI, and Jensen Huang of Nvidia. The document was titled the White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities. The President called it “morally binding.” The next day he signed an executive order directing federal agencies to call artificial intelligence “super intelligence.” The order says the new term covers exactly the same technologies as the old one.

It is tempting to dismiss the whole affair as theater. That would be a mistake. The accord is worth reading closely, because what it leaves out tells us more than what it says.

What it says

The text runs about 300 words. It asks each company building frontier AI models to adopt four layers of safeguards:

  • internal controls to monitor what their systems can do, with attention to cybersecurity, biological and chemical threats, and models that “hack or access technical systems in unintended ways”;
  • an internal team to confirm those controls actually work;
  • an independent outside auditor or evaluator;
  • an independent committee of the board of directors to receive the reports and make sure problems get fixed.

Anyone who has sat on a corporate board will recognize the structure. It borrows from how public companies police their financial statements. As governance design, it is sensible. If every frontier lab truly built all four layers and staffed them with people willing to say no, the world would be somewhat safer than it is today.

What it leaves out

The accord has no enforcement mechanism, no penalties, no funding, and no timeline. It does not say who the auditors will be, what they must test, or whether anyone outside the company will ever see their findings. It creates no regulator. It sets no definition of “safe.” Instead, it promises that the participating companies “will meet regularly to establish standards and best practices.” Those are the same companies whose products are being judged.

It also offers the companies no legal protection. There is no indemnification, no safe harbor, and no preemption of state law. Their legal exposure is exactly what it was the day before. But the accord ends by suggesting that “over time, it may make sense to codify these steps into laws or regulations.” House Speaker Mike Johnson has already floated it as a basis for legislation. The real question is what that future law would contain. A statute that turns these four voluntary steps into a liability shield or a ceiling on state regulation would be a very different thing from the accord itself.

We have seen this before. In July 2023 the Biden administration announced voluntary AI commitments from seven companies, several of them the same firms. Critics said then that no one could hold the companies to their word. Three years later, the record largely proved them right.

Who answers when something goes wrong?

Under this accord, the honest answer is: the companies answer to themselves. The board committee at the top of the structure is accountable to shareholders, not to the public. Citizens have no seat at the table, no access to the audits, and no remedy beyond the ordinary courts, which existed before the accord.

A compassionate libertarian reading

Those of us who distrust concentrated power usually aim our suspicion at the state. It is fair to note that this accord is not a government power grab. It imposes no new mandates and builds no new bureaucracy. Some readers will count that as a virtue.

But the libertarian case against concentrated power has never been only about government. It is about any arrangement in which a small number of actors hold great power while answering to no one. The accord creates exactly that arrangement and puts the White House seal on it. It is not the free market at work, because the public cannot comparison-shop between AI labs on catastrophic-risk safeguards it cannot see. It is not distributed, voluntary coordination among equals. It is six firms that together control most of the world’s frontier AI capability, certifying their own conduct, with the President as witness.

A genuinely liberty-minded framework would push power outward. It would require audit results to be made public, so customers, insurers, and investors can judge for themselves. It would let independent evaluators, not the companies alone, set testing standards. It would preserve the right of states, including Connecticut, which has debated its own AI legislation in recent sessions, to experiment with their own rules rather than having a voluntary federal pledge used as an argument to silence them. It would make whistleblower protection for AI workers explicit. None of this requires a new federal agency. All of it requires transparency, which the accord does not provide.

The case for the defense

To be fair, the accord has defenders, and their arguments deserve a hearing. Binding regulation of a fast-moving technology can freeze today’s assumptions into law and favor incumbents who can afford compliance departments. A voluntary framework can be revised as the science improves. Public commitments also carry some weight: a company that signed this document and then skipped the audits has handed future plaintiffs a ready-made exhibit. And not every signatory wants the same thing. Anthropic, for one, has publicly supported real safety legislation, while others in the room have lobbied against it. For some signatories the accord may be a floor, not a ceiling.

Those are reasonable points. But a floor that no one inspects is only a promise, and the public has been asked to accept such promises before.

The bottom line

The accord describes good practices. The companies deserve credit if they follow them. But in its current form it gives the most powerful firms of our time a government blessing, costs them almost nothing, and gives the rest of us nothing we can enforce. Before anyone turns it into law, citizens should insist on three things: audits we can read, auditors the companies do not choose alone, and no liability shields bought with voluntary pledges.

This post was written in collaboration with AI Matters, an editorial process combining human judgment with AI-assisted research and drafting. Disclosure: the AI system used in preparing this post is made by Anthropic, one of the accord’s signatories.

The Watching Crowd: What The Scarlet Letter Teaches About Power

Posted on by jdd

By JD DuRie, in collaboration with AI Matters

Americans tend to picture tyranny wearing a crown or a uniform. Nathaniel Hawthorne, writing in 1850 about Boston in the 1640s, offers a different picture, and it is more unsettling. In The Scarlet Letter, the most relentless power over a human life comes neither from the king nor, finally, from the magistrates. It comes from the neighbors.

The novel opens with a crowd outside a prison door. Hawthorne’s first observation is a dry one: whatever utopia of virtue a new colony’s founders imagine, they soon find it necessary to set aside one plot of ground for a cemetery and another for a prison. The Puritans came to New England seeking liberty of conscience, and within a generation they had built an apparatus for policing it. Hawthorne, whose ancestor John Hathorne sat as a judge at the Salem witch trials, knew this history from the inside.

Consider who speaks first when Hester Prynne is led out with her infant and her embroidered letter. It is not a minister or a governor. It is a cluster of townswomen, and they find the sentence too lenient. One suggests a branding iron. The ugliest of them says Hester has shamed them all and ought to die, and asks whether there is not a law for it. The magistrates, by comparison, look almost restrained. Hawthorne shows us something that the formal structures of government only partly explain: power in Puritan Boston is diffused through the watching crowd, and the crowd is often harsher than the law.

This is the novel’s most enduring insight for anyone who cares about liberty. Concentrated power is dangerous, and Hawthorne’s colony has plenty of it, with church and state fused together. But the scaffold works only because thousands of eyes are willing to stare. The letter is a government sentence, yet its real force is social: the children who follow Hester, the sermons that point to her, the strangers who fix their gaze on her breast. Tocqueville, visiting America fifteen years before the novel appeared, warned that democratic societies could produce a “tyranny of the majority” that works on the soul rather than the body. Hawthorne gives that warning a face.

What makes the book more than a protest, though, is how Hester responds. She does not flee, and she does not rebel in the obvious ways. She stays, sews, nurses the sick, and gives to the poor who sometimes curse her while taking her help. Over the years something remarkable happens: many townspeople stop reading the letter as its authors intended. They say it stands for “Able.” The community that imposed the meaning is gradually persuaded to revise it, not by a court ruling or a change in the law, but by one woman’s conduct over time. That is a distinctly voluntary, bottom-up kind of reform, and it is the most hopeful thing in the book.

Hawthorne is careful not to romanticize escape, either. In the forest, far from the settlement, Hester unpins the letter, lets down her hair, and tells her lover, the Reverend Dimmesdale, that what they did “had a consecration of its own.” She proposes they sail for Europe and begin again. It is the oldest American fantasy: the wilderness as a clean slate, freedom as simply walking away. The plan collapses. Dimmesdale instead climbs the scaffold and confesses before the town. Hawthorne suggests that liberty which ignores our bonds to other people is its own kind of illusion.

The novel also delivers a sharp moral ranking. Hester’s sin is public; Dimmesdale’s is hidden, and the concealment destroys him. The worst figure in the book is Roger Chillingworth, who has broken no law at all. His offense, in Dimmesdale’s words, is that he has “violated, in cold blood, the sanctity of a human heart,” treating another person’s inner life as a thing to be probed and possessed. In a society obsessed with outward conformity, Hawthorne locates the gravest wrong in the invasion of conscience. His closing moral is simple: “Be true! Be true! Be true!”

Readers may fairly object that this is a modern political reading of a book about guilt and passion. It is a reasonable challenge, and it deserves an honest answer. Hawthorne was no libertarian. He was a partisan Democrat, a friend and biographer of Franklin Pierce, and skeptical of reformers of every stripe, including those whose cause history has vindicated. Nor is his crowd purely villainous; it softens, it reconsiders, and Hester ultimately chooses to live among it again, wearing the letter by her own will and counseling troubled women who seek her out.

But those complications strengthen the reading rather than refute it. Hawthorne was not arguing for less government or more. He was showing that the question of freedom cannot be settled by constitutional design alone. A society can have liberty on paper and still crush the individual through gossip, surveillance, and the eager appetite for public shame. And the remedy he offers is not revolution but character: honesty in private, restraint in judgment, and the slow work of persuading one’s neighbors by example. There is irony here too. Hawthorne wrote the novel just after losing his post at the Salem Custom House when the political winds shifted, and his introductory sketch of that episode reads like a man who knew what it felt like to be judged by a faction.

For readers in 2026, the relevance hardly needs spelling out. Our scaffolds are digital, the crowd is global, and the letter can be pinned on someone in an afternoon. The machinery of public condemnation no longer needs a magistrate at all. Hawthorne’s warning is that a free people must watch not only those who hold office but also themselves when they gather to watch.

The Scarlet Letter is often assigned in high school and set aside afterward. It deserves a second reading as an adult, and particularly as a citizen. It is one of the earliest and still one of the wisest American books about the difference between a free society and a merely self-governing one.

This post was written in collaboration with AI Matters, an editorial process combining human judgment with AI-assisted research and drafting.

What Uncle Tom’s Cabin Still Asks of Us

Posted on by jdd

By JD DuRie, in collaboration with AI Matters

Harriet Beecher Stowe’s 1852 novel is, first and foremost, an argument about the moral horror of slavery — made not through abstraction but through the wreckage it leaves in Black families. Eliza fleeing across breaking ice with her son in her arms. Tom, sold away from his wife and children, then sold again, until he dies at the hands of Simon Legree rather than betray other enslaved people to save himself. Stowe’s case is built on personhood: that a system built on treating human beings as property is a standing contradiction of everything the Republic claims to believe about itself. That is the book’s center of gravity, and any reading of it has to start there.

But nested inside that larger argument is a narrower one, and it’s the one I find most durable for American civic life nearly two centuries later: Stowe’s insistence that law and moral obligation are not the same thing, and that when they diverge, conscience outranks statute.

She was writing in the shadow of the Fugitive Slave Act of 1850, which made it a federal crime — even in free states — to shelter or assist someone escaping slavery. Stowe’s answer, dramatized through the Quakers who shelter Eliza and the informal network we now call the Underground Railroad, was that ordinary people have not just a right but an obligation to defy a law that commands complicity in evil. That is a strand of the American tradition with real staying power — it runs backward through the founding-era case for a higher law than Parliament’s, and forward through Thoreau’s civil disobedience and the civil rights movement.

It’s worth being precise about what this argument is not, though, because it’s easy to over-read. Stowe was not a critic of state power as such. She wanted slavery ended by law — by federal force, if it came to that — and she supported the Union cause once war arrived. Her target was a specific unjust command, not government authority in general. The novel’s real claim is narrower and, I’d argue, more useful: that legality is not the same thing as legitimacy, and that citizens retain a moral accounting even when — especially when — the law tells them to look away.

It’s also worth holding the novel’s limits in view rather than past them. James Baldwin’s famous critique of Uncle Tom’s Cabin argued that Stowe substituted piety and sentiment for a genuine reckoning with slavery’s structural evil, reducing Tom to a symbol of patient suffering rather than a full person. The novel’s other Black protagonist, George Harris, complicates Stowe’s own preferred model — he argues for an active, even armed right of resistance, a harder and more assertive claim than Tom’s endurance. Reading the book only through the lens of quiet conscience risks repeating Stowe’s own softening of what resistance to injustice can require, and risks making the story’s white allies — the Quakers, the sympathetic Ohioans — its moral center, when its actual center is the personhood and judgment of the enslaved people themselves.

Held with those caveats, though, the case for individual conscience as a check on unjust law is the part of Uncle Tom’s Cabin that travels best. It doesn’t require you to accept a particular theology, and it doesn’t collapse into a brief for anarchy. It asks something more modest and more demanding: that citizens not outsource their moral judgment to whatever the statute book happens to say in a given decade. That’s an old American argument, and one worth re-reading in full — not just for its most portable lesson, but for the harder claims about personhood and resistance that the lesson too easily lets us skip past.

This post was written in collaboration with AI Matters, an editorial process combining human judgment with AI-assisted research and drafting.

Conscience, Incorporated — and the Machine That Came After

Posted on by jdd

A review of Conscience Incorporated: Pursue Profits While Protecting Human Rights by Michael H. Posner (NYU Press, 2024)

Some books are worth reading precisely because they are wrong in instructive ways. Michael Posner’s Conscience Incorporated is not one of them. It is worth reading because it is largely right about the disease and only half-right about the cure — and the gap between those two halves is where the most important argument of our moment is being fought.

Posner is no bomb-thrower. He is the Jerome Kohlberg Professor of Ethics and Finance at NYU’s Stern School of Business, founder and chair of the Fair Labor Association, and Assistant Secretary of State for Democracy, Human Rights, and Labor under President Obama. He has spent a career in the room where the compromises get made, and the book carries the calm authority of someone who has watched a lot of corporate promises curdle. His thesis is bracingly simple: the old story that pits profit against ethics is a false binary, and the firms that thrive over the long run will be the ones that build human-rights diligence into their core operations rather than quarantining it in a corporate social responsibility department that exists to issue glossy reports. He builds the case through familiar defendants — Nike, Coca-Cola, Walmart, Meta — and argues, persuasively, that public awareness alone has never been enough to discipline a global supply chain.

I want to start by giving the man his due, because there is more common ground here than a libertarian reader might expect.

What Posner gets right

A compassionate libertarian and a human-rights lawyer should agree on one foundational point: concentrated power is dangerous to the people on the wrong end of it, and it tends to escape accountability unless something forces the issue. Posner’s catalog of abuses — the contractor four time zones away whose labor never appears on an org chart, the moderator absorbing the worst of the internet for a subcontractor’s wage, the supplier whose conditions the brand prefers not to inspect too closely — is not a fantasy. These are real harms done to real people, and a philosophy that takes individual dignity seriously cannot wave them away with an invisible hand and a shrug.

Better still, Posner’s central operational insight is, in spirit, anti-bureaucratic. Embedding diligence into how a company actually buys, builds, and ships — rather than bolting on a compliance theater — is exactly the kind of bottom-up, incentive-aligned reform that should appeal to anyone skeptical of box-checking. When he frames human-rights risk as business risk — reputational, operational, eventually financial — he is speaking the market’s own language. A firm that treats its workers and users as means to an end, and gets caught, pays for it. That is not a regulatory truth; it is a competitive one.

So far, so good. The trouble begins when Posner reaches for the lever.

The machine he didn’t fully see coming

The book went to press just as the ground shifted. Artificial intelligence does not break Posner’s diagnosis — if anything, it vindicates it — but it stress-tests his prescription nearly to failure.

Consider his core concern, the worker rendered invisible by distance and subcontracting. AI has manufactured an entirely new version of that worker: the data labeler. The large models now reshaping white-collar work were trained, in part, by a “tasker underclass” in the Global South — workers in Kenya, Uganda, and India paid on the order of thirteen to sixteen dollars a day to sort, rate, and sanitize the raw material of machine intelligence, including the disturbing content the rest of us never have to see. This is Posner’s chapter on outsourced labor, rewritten for a new decade, and it is arriving faster than any framework designed for garment factories can track.

At the same time, the displacement has climbed the ladder. The anxiety is no longer confined to the loading dock; it has reached the analyst’s cubicle and the junior associate’s desk. The fear among serious people is not science fiction — the chief executive of one leading AI lab has warned that the technology could eliminate a large share of entry-level white-collar jobs and push unemployment into the double digits. Whether or not that forecast proves accurate, the human stakes Posner cares about have expanded, not contracted.

And on the question he spends real ink on — platform content and online harm — AI now sits on both sides of the ledger. It lets companies moderate at a scale no human staff could match, and it spares flesh-and-blood moderators some of the trauma of the job. It also amplifies the bias baked into its training, hallucinates with confidence, and hands every troll and propagandist a generative firehose. The technology is simultaneously the best tool we have for the problem and a powerful new source of it.

In other words: the moral case Posner makes is more urgent in 2026 than it was in 2024.

Where the libertarian parts company

But Posner’s instinctive remedy is government, and his preferred model is Europe — mandatory due-diligence regulation, enforced from above. Here the compassionate libertarian has to register a respectful dissent, and AI is the cleanest illustration of why.

First, the pacing problem. Regulation moves at the speed of committee; AI moves at the speed of capital. In the time it takes a legislature to define a harm, the technology that produces it has been superseded twice. This is not a hypothetical complaint. The United States has spent the period since the book’s publication moving away from the regulatory approach — a December 2025 executive order now directs the Justice Department to identify and challenge state AI laws and to assemble a federal framework explicitly designed to preempt them. Europe, the model Posner holds up, has proposed delaying the high-risk obligations of its own AI Act from 2026 to 2027. The remedy he prescribes is in political retreat on both continents precisely because the thing it was meant to govern outran it.

Second, the capture problem. Mandatory regulatory regimes are written, in practice, by the largest incumbents — the only players with the lawyers and lobbyists to be in the room. A compliance burden that a trillion-dollar platform absorbs as a rounding error can be a moat that drowns the smaller competitor who might otherwise have done the work better. Rules sold as a check on corporate power have a way of entrenching the most powerful corporations of all. Posner is admirably suspicious of concentrated corporate power; he is notably less suspicious of concentrated state power, even though the historical record gives ample reason to fear both.

Third, and most fundamentally: the book is more confident than I am that the conscience belongs in the statute book. The mechanisms that have actually moved corporate behavior — investors who reprice risk, customers who walk, employees who refuse to build the thing, journalists and researchers who drag the subcontract into daylight — are distributed, voluntary, and fast. They are also, not incidentally, the mechanisms Posner himself spent a career using to good effect before he concluded the answer was a mandate. The Fair Labor Association is a voluntary association. The shaming of bad actors is a market and civil-society process. The most effective tools in his own toolkit are the ones that don’t require a regulator at all.

An honest accounting

None of this makes Conscience Incorporated a bad book. It makes it an important and incomplete one. Posner has correctly identified that scale plus distance plus opacity equals abuse, and that “the market will sort it out eventually” is cold comfort to the person being harmed right now. That is a serious moral point and it deserves a serious answer.

But the answer, I’d argue, is not to transfer the conscience to the one institution with a monopoly on force and a documented record of being captured by the very interests it claims to police. The answer is to make the distributed mechanisms sharper — radical transparency, genuine liability for genuine harm, the freedom of workers and users and investors to exit and to expose. Treat people as ends rather than means, yes. Posner and I agree on the destination. We disagree about who should be holding the map, and AI has made that disagreement the central economic question of the decade rather than a footnote to it.

Read the book. Argue with the second half. The first half will have already won you over, and the argument over the rest is the one worth having.

JD DuRie writes on Connecticut civic life, political philosophy, and the long argument between liberty and power at CT Capital. This review was developed in collaboration with AI Matters.