Launching the Institute for AI and Democratic Governance
From Blindness to Clarity
This photograph was taken outside UNESCO House in Paris, where we attended the 2026 meeting of the International Association for Safe and Ethical AI. I am on the left. Beside me is Chris West, the Institute's Development Director and a founding partner in this work. It seemed right to launch the Institute for AI and Democratic Governance from a place where the question animating our work — who will write the rules of the algorithmic age, and whether those rules will be written by democratic peoples or for them — is already being asked at the highest levels of international deliberation.
Those of you who have followed this publication know that my last essay announced my departure from the legal academy — an institution I had served for more than twenty years, and which I left not in bitterness but in what I can only describe as an act of fidelity. Fidelity to the work. Fidelity to the students. Fidelity to the belief that democracy is a moral practice, not a market function, and that legal education exists to form the kinds of persons democracy requires.
Today I want to tell you what comes next, and why.
The Problem Is Blindness.
I want to be precise about something, because the precision matters both intellectually and personally. The failure of the legal academy at this moment is not, in the main, a failure of intention. The deans, accreditors, faculty committees, and administrators who are presiding over the hollowing out of legal education’s democratic mission are not, for the most part, cynics or villains. They are people working within frameworks that were built for a world that no longer exists, and those frameworks have made them structurally incapable of seeing what needs to be seen.
This is not a new diagnosis. Anthony Kronman, in The Lost Lawyer, traced the long decline of what he called the lawyer-statesman ideal (the vision of the lawyer as someone who combines technical legal mastery with practical wisdom, civic virtue, and a genuine commitment to the public good). That ideal, Kronman argued, once animated both the practice of law and its pedagogy. It understood that legal education was more than merely professional preparation, but also includes the deeper work of civic formation. Law schools were once engaged in the cultivation of persons capable of sustaining democratic institutions through the exercise of deliberative judgment. This was the meaning of being a “public citizen.” It has been progressively abandoned, displaced by the rise of law and economics, the dominance of large firm culture, and the transformation of legal education into a technical skills-oriented enterprise more concerned with the commercial exploitability of their “product” (students) than moral formation of public citizens.
The Carnegie Foundation’s landmark report on legal education, published in 2007, confirmed the damage from a different angle. Legal education, the Carnegie Report found, excels at teaching students to think like lawyers — to analyze, to argue, to move within the formal structures of doctrine. What it systematically neglects is the formation of professional identity, practical wisdom, and the ethical commitments that make lawyering a vocation rather than a trade. The academy saw this diagnosis clearly. It was discussed at conferences and in faculty meetings. And then the financial crisis of 2008 intervened, and what might have been a moment of institutional renewal became a moment of institutional retrenchment instead.
The collapse of the legal employment market through 2009 and 2010 produced a panic that hardened into an identity. Law schools, whose reputations were increasingly tied to employment outcomes, turned decisively toward commercial values (bar passage rates, employer relationships, market placement) and away from the formative agenda Carnegie had urged. This was not cynicism. It was a survival response that became a worldview. By the time the market stabilized, the new logic was entrenched. The administrator who later asked whether we might teach “the ethics that employers want” was not departing from his institution’s mission. He was faithfully executing it as 2008 had redefined it.
The result was a legal academy that greets artificial intelligence as primarily a skills problem. It is treated as a literacy challenge, a set of tools to be mastered, a source of new competencies for an anxious market. It is not seen as a transformation of governance that requires exactly the kind of deliberative moral judgment the academy has spent thirty years systematically deprioritizing. The academy is, in a phrase that has stayed with me, fighting over the architecture of a building whose foundation is shifting. The accreditation debates, the AI literacy initiatives, the legal tech centers — all of it is rearranging deck chairs. The deeper question, the one that would require genuine institutional transformation to address, goes not just unanswered but unasked.
That is the blindness. And it is now dangerous.
What the Blindness Costs
We are at a threshold moment in the development of artificial intelligence . We are moving from systems that recommend to systems that act. This transition is already creating a crisis of accountability that existing legal, ethical, and institutional frameworks are not equipped to address.
Automated systems now make or heavily influence decisions about criminal sentencing, credit allocation, hiring, healthcare rationing, benefit eligibility, and civic participation. These decisions are optimized for efficiency, consistency, and computational tractability. They are precisely not optimized for justice, dignity, or the kind of contextual moral reasoning that democratic governance has historically required.
The defining political condition we are entering is not characterized by AI that malfunctions. It is characterized by AI that works.
I call this the Agentic State: a form of governance that achieves technical perfection at the cost of ontological emptiness. It produces outcomes but extinguishes the political conditions of deliberation, contestation, accountability, moral address that make those outcomes democratically legitimate. Hannah Arendt diagnosed something related in a different context: the replacement of action by behavior, the substitution of calculable, rule-governed process for the unpredictable, irreducible agency of persons acting together in public life. What she saw as a political danger is now being institutionalized at an industrial scale, in systems that process millions of decisions per day, none of which addresses the person before it as a person.
It is worth pausing on that last phrase. None of which addresses the person before it as a person. This is not a technical malfunction. It is the system working precisely as designed. The merchant, to borrow a formulation I find clarifying, knows the price of everything and the taste of nothing. The Agentic State knows the risk score of everyone and the moral weight of nothing.
Here is where the legal academy’s blindness becomes most costly. The legal profession occupies a distinctive position in a democratic society. Tocqueville called it the country’s democratic aristocracy — the professional class whose civic commitments and institutional independence serve as a check on both popular excess and executive overreach. Lawyers are not primarily service providers. They are officers of the court, advocates for rights, interpreters of constitutional meaning, the people who stand between citizens and the institutions that govern them. The lawyer-statesman ideal that Kronman mourned was not nostalgia for a golden age. It was a description of what law requires if it is to remain a democratic institution rather than a technical system.
That function is now under simultaneous pressure from two directions. From the market side, AI is bifurcating the profession rapidly: elite lawyers will increasingly orchestrate AI systems, while routine legal work disappears. The middle of the profession (solo practitioners, public defenders, legal aid attorneys, the lawyers most connected to ordinary people’s access to justice) is being hollowed out. From the institutional side, the systems that credential lawyers are philosophically unprepared for what is happening. Legal education was designed for a stable conception of what law is, what lawyers do, and what role the profession plays in democratic society. AI is destabilizing all three simultaneously.
The same historical moment in 2008 that pushed the legal academy toward commercial values also accelerated the financialization of algorithmic decision-making: credit scoring, risk assessment, automated underwriting, predictive policing. The academy and the Agentic State were shaped by the same historical forces. And the academy’s commercial turn left it structurally unable to respond to the governance crisis that those same forces produced. It cannot ask the formative question — what kind of persons does democracy require, and are we forming them? — because 2008 taught it to ask a different question instead: what kind of graduates does the market want, and are we producing them? These are not the same question. They have not produced the same lawyers. And the difference is now consequential in ways that can no longer be ignored.
The Logical Foundation
Before I describe what the Institute does, I want to give you the philosophical core because if the argument is right, everything else follows, and if it is wrong, nothing else matters.
Drawing on Jürgen Habermas’s insight in Between Facts and Norms, I argue that autonomous systems are constitutively incomplete. Algorithms can optimize for any objective function they are given. But they cannot determine whether that objective function is just, democratically authorized, or adequate to the moral complexity of the situation it governs. Legitimacy cannot be computed; it must be earned in the hearts of the governed. And, justice cannot be automated because to do so would preclude the purpose of justice, which is to recognize the dignity of the person.
This means that human judgment in high-stakes decision environments is not merely a preference to be weighed against efficiency gains. It is a logical necessity. This gives us something much stronger than an ethical preference. It provides a principled, formal foundation for insisting on human judgment that can move from seminar rooms into courtrooms and regulatory proceedings.
From this foundation, I have developed what I call ontological due process, which isthe constitutional and moral argument that every person who stands before an automated decision-making system retains the right to be treated as a person rather than a managed datum. This is not simply a procedural claim about notice and hearing. It is a foundational claim about what law is for. Ontological due process entails three concrete protections. The right to a moral interlocutor: to have one’s case heard by a human being capable of moral reasoning and accountable for the exercise of judgment. The right to narrative contestation: to present one’s life as a story rather than a data profile, to introduce context that no training set can capture, and to challenge the adequacy of the categories through which one is being judged. And the right to inefficiency: the insistence that the friction and vulnerability of human deliberation is not a defect to be optimized away but the constitutive condition of democratic legitimacy. It is precisely in the space of deliberation, error, and accountability that democratic life is produced and sustained.
If Jean Bethke Elshtain was right (and I believe she was), democracy is not merely a mechanism for aggregating preferences. It relies on the vital, contested, and deeply moral spaces of civil society. It relies on the capacity of citizens to engage in shared deliberation, to be addressed as moral agents rather than managed as data profiles. The three protections of ontological due process are not merely legal claims. They are descriptions of what democratic personhood requires in order to survive contact with the Agentic State.
These ideas are developed at length in a forthcoming article in the St. Thomas Journal of Law and Public Policy and anchor a monograph trilogy I am completing.
What the Institute Does
The Institute for AI and Democratic Governance exists to close the distance between knowing what is right and building the institutions that require it. It is the translation problem — from ethical “should” to enforceable institutional design — that has defined my career and that no existing institution is adequately addressing.
We work on three levels simultaneously, because the crisis operates on three levels simultaneously. At the level of hard law, we apply specific legal doctrines — Caremark fiduciary duties, professional negligence standards, agency law — to hold institutions accountable for the AI systems they deploy. If a corporate board delegates oversight to a black-box algorithm, who is liable? If an AI agent causes harm, where does accountability rest? These are not theoretical questions. They are doctrinal ones, and they require the kind of legal expertise that most AI ethics discourse conspicuously lacks. This is the plumbing of governance: unglamorous, technical, and indispensable.
At the level of the diagnostic framework, we bring the full resources of legal theory, continental philosophy, and democratic thought to bear on what is actually changing in the structure of governance itself. The Habermas argument, the Arendt-Foucault-Agamben genealogy of the Agentic State, the personalist tradition of Elshtain — these are not decorative references. They are the analytical tools that allow us to see what is happening with the clarity that existing frameworks cannot provide. You cannot apply old laws to new code without understanding what has changed in the legal system itself. That understanding is what the Institute exists to produce and disseminate.
At the level of moral stakes, we insist on the formative question that the commercial academy has abandoned. We evaluate AI systems and governance proposals not by computational efficiency but by their capacity to expand rather than constrict human potential — to preserve the conditions of moral purpose that democratic personhood requires. And we build educational programs designed to cultivate what Elshtain would call democratic formation: the civic capacity to act as a moral agent in institutions increasingly designed to make moral agency unnecessary.
This includes continuing legal education curricula for practicing lawyers grounded in hard doctrinal skills rather than mere ethical aspiration. It includes graduate course development and high-level convenings that bring technologists, legal theorists, and democratic philosophers into sustained dialogue. And it includes a public-facing online course program based on my forthcoming book, Our Unfinished Work: Reclaiming Democracy in the Age of AI (Johns Hopkins University Press), designed to reach the lawyers, civic leaders, educators, and citizens who sense that something is profoundly wrong with where AI is taking democratic life but lack the vocabulary to name it precisely.
Ontological due process is that vocabulary. The Institute exists to put it to work.
Why Outside
I want to be direct about one more thing, because you deserve honesty rather than institutional posturing.
The Institute is a standalone nonprofit, independent of any university. This is not a limitation I am managing. It is a structural choice that the analysis demands.
An institute housed in a law school would face the same institutional grammar that made the work unwelcome in the first place. It would be subject to the commercial logic that 2008 institutionalized. It would be accountable to the market pressures that teach administrators to ask for the ethics employers want rather than the ethics democracy needs. It would be captured by the very blindness it exists to address.
Kronman ends The Lost Lawyer with a question rather than an answer: whether the lawyer-statesman ideal can be recovered from within institutions that have so thoroughly abandoned it. My experience across more than twenty years in those institutions has persuaded me that the honest answer, at least for the work this moment requires, is no. Not because the people within them are incapable of seeing, but because the frameworks within which they operate have made certain questions illegible. The question of democratic formation, of what kind of persons law schools should be producing and why democracy depends on the answer, is not legible within a framework that measures success by employment outcomes and bar passage rates. It requires a different institutional home.
Independence is the precondition for clarity. Exile, as I wrote when I launched this publication, affords a contrapuntal view — the ability to see things as they are rather than as institutional belonging requires us to see them. The Institute is the institutional form of that view. It is the attempt to build, from outside the walls that have forgotten what they are for, something adequate to what this moment requires.
That is not bitterness. I want to be clear about that. The people who could not see the work I was bringing are not enemies. They are people working within frameworks that have made them blind. Some of them are good people. Some of them taught me things I still carry. What I owe them is not grievance but the honest work of building something they could not — not to defeat them, but because the moment demands it.
An Invitation
Democracy has always been, in Lincoln’s words, an unfinished work. Its survival has never depended on the perfection of its institutions but on the willingness of persons — and the institutions they build — to keep the work going. The question AI poses to our generation is whether the next chapter of that unfinished work will be written by engaged citizens exercising democratic judgment, or by systems optimized for everything except the things democracy requires.
Closing the distance between those two possibilities is the defining work of this moment.
I founded the Institute because I believe that work requires an institution built for independence, philosophical seriousness, and the slow, unglamorous labor of building governance adequate to the crisis. I am under no illusion that this will be easy. I am under no illusion that it will be fast. But I am convinced it is necessary, and I am convinced that the readers of The Exile — people who have chosen, in their own way, to see clearly rather than comfortably — are exactly the community this work needs around it.
I welcome the conversation. I welcome the challenge. I welcome the company.
The Exile continues. The work is just beginning.