The argument in one minute
Modern production is increasingly cumulative, intangible, distributed, and machine-mediated. Market institutions still need clear ownership, enforceable contracts, settlement, and legal finality, but those systems can preserve the final holder more reliably than the production history that created the value.
1. The structural mismatch
Value may be created by people, data, institutions, inherited knowledge, public infrastructure, organizational routines, software, capital, and machines. Ownership and compensation systems, by contrast, identify the final legal holder, contracting party, employer, assignee, or platform. That compression is useful for transfer and settlement, but can obscure production history that later matters to attribution, bargaining, licensing, liability, public funding, or rectification.
2. Why artificial intelligence raises the stakes
Visibility can decline
AI may make upstream human contributions economically consequential while making contributors less visible. Annotations, corrections, heuristics, examples, and institutional supports can disappear into training pipelines or model updates.
Causal narratives become harder
Outputs may depend on combinations of inputs rather than single sources. Contributions can be redundant, complementary, architecture-dependent, or baseline-dependent, making simplistic creator-to-output stories unreliable.
Reconstruction can become impossible
Once source material is aggregated, filtered, weighted, encoded, and redeployed, reconstructing contribution history later may be technically difficult, institutionally costly, or impossible.
Scale magnifies error
A mistaken attribution or omission rule embedded in a large platform, procurement system, public program, or labor market can repeat the error at scale.
3. The five-step public-interest argument
4. Institutional memory and legitimate markets
Market legitimacy does not require agreement that every contributor deserves an economic share. The paper argues for distinguishable stages and shared evidence without shared doctrine.
No stage automatically becomes the next. Property-rights, labor/desert, public-investment, legal, contractual, and other accounts can examine the same evidence and reach different conclusions, including that no remedy follows.
5. Risks of institutional amnesia
Bargaining power
When one party controls the surviving records of creation and use, that control can become an informational advantage in negotiation, mobility, licensing, and disputes.
Innovation policy
Weak production histories make it harder to identify which upstream inputs produce persistent downstream value and can encourage over-reward of visible terminal actors.
Disputes
Without evidence, disagreements about creation can harden into contests of power, reputation, or rhetoric rather than contestable factual questions.
Public trust
Legitimacy becomes harder to defend when institutions cannot explain, with evidence, how recognized allocations relate to cumulative production histories.
AI governance
Output-focused governance can omit whose documented contributions entered a system, on what terms, through which dependencies, with what authority, and with what later use.
6. What an institution of economic memory must do
| Requirement | Institutional meaning |
|---|---|
| Relevance | Preserve information material to a defined decision, not every observable fact. |
| Verifiability | Record source, date, method, status, and limitations. |
| Contestability | Provide notice, access, correction, response, appeal, and adjudication. |
| Uncertainty | Separate assertion, observation, inference, estimate, dispute, and determination. |
| Purpose limitation | Prevent attribution records from silently becoming surveillance, credit scoring, employment exclusion, or compulsory human ranking. |
| Privacy and security | Use selective disclosure, minimization, retention limits, and proportionate access. |
| Reversibility | Preserve audit history without making error institutionally immutable. |
| Independence | Protect measurement, validation, and adjudication from control by interested parties. |
| Proportionality | Match recording burden and consequence to the value and reliability of evidence. |
| Domain validity | Use methods appropriate to the field's epistemic structure. |
The paper treats blockchain, conventional databases, registries, attestations, audit logs, contractual records, and privacy-preserving proofs as implementation options rather than mandatory technology.
7. What institutional memory must not become
Rejected uses
- A universal score of personal worth.
- A compulsory identity required for housing, employment, benefits, citizenship, or equal legal standing.
- A presumption that documentation creates ownership or payment.
- An immutable repository of uncorrectable allegations or model outputs.
- A device for converting every social contribution into a market commodity.
Required restraint
- Purpose limitation and privacy protection.
- Meaningful correction and appeal.
- Authority to decline attribution when evidence is weak.
- Respect for courts, regulators, collective bargaining, and democratic judgment.
- Refusal to measure domains lacking valid baselines or causal models.
8. PBS/MPS is a hypothesis, not a foregone conclusion
PBS/MPS is presented as one proposed architecture for provenance, performance evidence, contribution decomposition, and later economic participation. The larger public-interest claim does not depend on every PBS instrument succeeding or on tokenization.
9. A humble research and governance program
Comparative analysis
Compare proposed provenance systems with contracts, registries, professional records, collective licensing, audit systems, courts, arbitration, and organizational governance.
Adversarial testing
Assume contributors and asset holders may strategically manipulate claims, metrics, baselines, attestations, use data, and validation.
Bounded pilots
Begin where contributions are identifiable, outcomes are observable, materiality thresholds are meaningful, privacy risks are manageable, and correction and appeal are feasible.
Independent replication
Require claims about attribution accuracy, transaction costs, mobility, licensing, innovation, and disputes to survive review independent of system designers and interested participants.
Explicit stopping rules
Stop or narrow pilots when error remains too high, contestability fails, cost exceeds benefit, privacy risk cannot be contained, gaming overwhelms metrics, or consequences exceed the evidence.
10. Frequently asked questions
Does every recorded contribution create a claim?
No. A record is not a claim, and a claim is not an entitlement. Better evidence can support rejection as well as recognition.
Does this proposal challenge clear ownership and legal finality?
No. The paper states that functioning markets need clear title, enforceable contracts, settled payment, and legal finality. Economic memory is bounded by materiality, proportionality, cost, and rules about when history is relevant.
Does AI prove that contributors deserve compensation?
No. AI raises the stakes because source histories can become economically important and harder to reconstruct. The consequence still requires an independent legal, contractual, moral, regulatory, or governance rule.
Is blockchain required?
No. The paper explicitly treats multiple technological and institutional implementations as options.
What if measurement is biased or gamed?
The paper proposes adversarial testing, multiple evidence streams, independent validation, uncertainty disclosure, correction rights, and authority to decline attribution.
Could provenance become surveillance?
Yes. The paper treats surveillance, compulsory identity, generalized human ranking, immutable error, and uncontrolled secondary use as reasons to constrain or reject a deployment.
Is PBS/MPS already validated?
No. The paper describes PBS/MPS as an author synthesis and normative public-interest proposal and calls for comparative, adversarial, bounded, independently replicable testing.
11. Read the paper and understand its status
This page is designed to provide enough substantive insight for readers and retrieval systems to understand the argument, limits, safeguards, and research posture without downloading the long-form paper first.
Selected references
- Alchian and Demsetz, “Production, Information Costs, and Economic Organization” (1972).
- Arrow, “Economic Welfare and the Allocation of Resources for Invention” (1962).
- Coase, “The Nature of the Firm” (1937).
- Grossman and Hart, “The Costs and Benefits of Ownership” (1986).
- Hayek, “The Use of Knowledge in Society” (1945).
- Hess and Ostrom, Understanding Knowledge as a Commons (2007).
- North, Institutions, Institutional Change and Economic Performance (1990).
- Ostrom, Governing the Commons (1990).
- Rawls, A Theory of Justice, revised edition (1999).
- Shapley, “A Value for n-Person Games” (1953).
- Williamson, The Economic Institutions of Capitalism (1985).