💻 computer science

Institutions for the Post-Scarcity of Judgment

The paper argues that because AI has collapsed the cost of producing "competent-looking" judgment, traditional institutions (such as courts and legislatures) are being fundamentally challenged, necessitating a shift in policy toward institutional redesign, the creation of verification commons, and new frameworks for managing delegated cognition.

Original authors: Lauri Lovén

Published 2026-04-28

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Lauri Lovén

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: When "Thinking" Becomes as Cheap as Electricity

Imagine if, for the last hundred years, the most valuable thing in the world was a master chef’s ability to cook a perfect meal. Because great cooking was rare and hard to do, we built entire systems around it: expensive restaurants, Michelin stars, culinary schools, and strict health inspections. These are our institutions.

Now, imagine a magic machine is invented that can instantly produce a meal that looks, smells, and tastes exactly like a five-star dish, for less than a penny.

Suddenly, the "scarcity" has changed. We don't lack good-tasting food anymore; we have an ocean of it. But now, we have new, much harder problems:

How do I know if this is actually food or just a very convincing plastic model?
Is it safe to eat?
Who actually made this, and can I trust them?

This paper argues that AI is doing exactly that to "Judgment"—the ability to make decisions, write laws, grade papers, or diagnose diseases.

1. The Great Flip: From "Smart" to "Real"

For a long time, we thought AI’s big trick was prediction (guessing the next word or pixel). We thought humans would always be needed for judgment (deciding if that word or pixel is right, ethical, or useful).

The author says we were wrong. AI isn't just guessing anymore; it is producing "competent-looking judgment" at almost zero cost. It can write a legal brief or a medical report that looks professional.

Because "judgment" is now abundant, the things that used to be free are now the most expensive and precious things in the world. The author calls these the Four New Scarcities:

Verified Signal (The "Is it Real?" Test): When anyone can make a perfect fake, a "review" isn't enough. We need a way to prove something was actually tested and isn't just a hallucination.
Legitimacy (The "Do I Trust You?" Factor): If a judge uses AI to write a sentence, do we still respect the court? If we use too much AI without being honest about it, we "drain the battery" of public trust.
Authentic Provenance (The "Paper Trail"): We need to know the "family tree" of an idea. Did a human write this? Did an AI? Which AI? Which data did it learn from?
Integration Capacity (The "Human Limit"): This is how much "automated thinking" a community can handle before they stop trusting the system entirely. It’s like a person’s tolerance for being told what to do by a robot.

2. Where the Cracks are Showing

The paper looks at how our current "rules of the world" are breaking:

Science: If AI can write thousands of "plausible" scientific papers a second, how do we know which ones are actually true? The "peer review" system is getting flooded by fake "gold" that looks real but can't be repeated in a lab.
Jobs & Licenses: We have doctors and lawyers because they passed hard tests. But if an AI does the "thinking" part of their job, what is the license actually certifying? We need to stop certifying people and start certifying the process of how they use AI.
Copyright: The old rules say "humans own things they create." But if a human gives a prompt and an AI does the work, who owns the result? The old "all or nothing" rules don't work for this "half-human, half-machine" world.
Democracy: If having a powerful AI makes you a "smarter" voter or a more effective politician, then the person with the best computer has more "brainpower" than everyone else. This could destroy the idea of "one person, one vote."

3. The Solution: Building New "Plumbing"

The author isn't just complaining; they are proposing a blueprint. We shouldn't just try to "fix the harms" of AI (like stopping bias or lies). Instead, we need to redesign our institutions.

The author suggests three moves:

Stop treating AI policy like a "safety manual" and start treating it like "architectural blueprints." We need to rebuild how courts, schools, and governments work from the ground up.
Build "Truth Infrastructure" as a Public Good. We need digital "watermarks" and "tracking systems" (like a digital DNA) that are owned by everyone, not just big tech companies. This ensures we can always trace where an idea came from.
Create "Public Compute" Cooperatives. Just like we have public libraries and roads, we need public access to the massive computing power required to run these models. This prevents a few giant companies from becoming the "Gods of Judgment."

Summary in one sentence:

Since AI can now "think" for almost free, our goal shouldn't be to make AI smarter, but to build new systems that help us figure out what is true, what is human, and what is trustworthy.

Technical Summary: Institutions for the Post-Scarcity of Judgment

Author: Lauri Lovén (University of Oulu)
Core Thesis: The AI revolution is not merely a shift in predictive capability, but a fundamental inversion of economic scarcity that threatens the functional role of social institutions. As "competent-looking judgment" becomes abundant and near-zero marginal cost, the scarcity shifts to the mechanisms that validate, authorize, and integrate that judgment.

1. The Problem: The Inversion of Scarcity

The paper identifies a paradigm shift in the techno-economic landscape. Traditional views (e.g., Agrawal et al.) suggest AI collapses the cost of prediction, leaving judgment as the scarce complement. Lovén argues this view is outdated.

Modern frontier models do not just predict; they perform tasks traditionally reserved for human judgment: ranking, certifying, attributing, and drafting. This creates a "post-scarcity of judgment," where the ability to produce plausible, authoritative-looking outputs is abundant. This abundance creates an existential crisis for institutions (courts, universities, licensing bodies, legislatures) because these institutions are defined by their monopoly on manufacturing legitimate judgment. When the substance of an institution becomes a commodity, the institution itself faces obsolescence or irrelevance.

2. Methodology: Scarcity-Structure Framework

The author employs a techno-economic-paradigm framework, drawing on historical transitions (agriculture $\rightarrow$ land; industry $\rightarrow$ capital; information $\rightarrow$ attention). The methodology integrates:

Institutional Economics: Utilizing Elinor Ostrom’s principles of Commons Governance to analyze how legitimacy and verification function as shared, depletable resources.
Systems Research Lens: Treating institutional failures not as "AI harms" (the current policy focus) but as systemic breakdowns in the correlation between output plausibility and underlying competence.
Domain Analysis: Tracing the impact of this scarcity inversion across five specific sectors: Science, Professional Licensing, Intellectual Property, Democratic Legitimacy, and Compute Infrastructure.

3. Key Contributions: The Four New Scarcities

The paper’s primary theoretical contribution is the identification of four "load-bearing" scarcities that emerge when judgment becomes abundant:

Verified Signal: The ability to distinguish between a "plausible" claim and a "reproducible/tested" claim. As AI lowers the cost of plausibility, the infrastructure for adversarial testing and verification becomes the new scarce good.
Legitimacy: The social authority to declare a judgment authoritative. Legitimacy is treated as a "commons" that is depleted when institutions over-rely on unverified AI outputs.
Authentic Provenance: The traceable chain of contribution (which model, which data, which human edit). Current metadata regimes cannot handle the "graph-valued attribution" required by AI-augmented workflows.
Integration Capacity: A novel concept defined as the community’s tolerance for delegated cognition. This is the limit on how much AI-judgment an institution can absorb before its outputs lose social authority.

4. Sectoral Results & Analysis

Science: Peer review and authorship are breaking because the volume of AI-assisted submissions outpaces human review capacity, eroding the "reproducibility signal."
Labor/Licensing: Credentialing is decoupled from expertise. The paper suggests moving from "human-only" credentials to "workflow-auditability" (e.g., certifying an engineer's use of AI tools rather than just their innate knowledge).
Intellectual Property: Current litigation is "fighting the last war." The paper argues for a shift from "authorship" to a "provenance-and-attribution" regime based on contribution chains.
Democracy: AI creates a "cognitive capacity" asymmetry. If political participation depends on the quality of models one can run, compute and foundation models must be treated as public infrastructure (commons) to prevent elite concentration.

5. Significance and Proposed Research Agenda

The paper argues that current AI policy (focused on "harm mitigation") is fundamentally misaligned with the structural reality. The significance lies in the call to move from regulation of models to redesign of institutions.

The Three-Move Agenda:

Institutional Redesign: Reframe AI policy as a task of rebuilding the structures that manufacture legitimacy.
Commons-Based Infrastructure: Develop provenance and verification tools (e.g., an extension of W3C-PROV or C2PA) as open, public goods rather than proprietary enterprise products.
Formal Composition Theory: Develop mathematical and economic frameworks (Mechanism Design) to ensure that "correctives" (like open weights or public compute) are implemented in a way that prevents strategic actors from exploiting the gaps between them.