What Is Fairness? Why AIIM Could Become Humanity’s Fairest Engine for Distributing Money

Getting your Trinity Audio player ready...

Fairness means applying the same relevant criteria to ourselves that we apply to other people. If we regard our own need for food, security, recognition, meaningful work, or financial support as important, consistency requires us to recognize comparable needs in others. Fairness does not necessarily mean giving everyone the same amount. It means that differences in treatment must follow relevant, publicly defensible principles rather than status, personal connections, prejudice, or institutional power.

This principle is simple to state but difficult to implement. Human beings usually understand their own circumstances in detail while seeing other people only through incomplete records, social labels, and institutional reputations. As a result, money is often distributed according to visibility, credentials, geography, persuasive ability, and access to decision-makers—not according to either genuine need or useful contribution.

AI Internet-Meritocracy (AIIM) is a proposed alternative. It is designed to evaluate publicly documented scientific and open-source contributions and distribute available funds proportionally, continuously, and transparently. Its central promise is not that an AI model is automatically fair. Rather, AIIM attempts to turn fairness from an inconsistent personal judgment into an auditable social procedure.

A Practical Definition of Fairness

Fairness can be summarized as follows:

Relevant similarities should be treated similarly, and different treatment must be justified by relevant differences.

Suppose two researchers make contributions of comparable importance. If one receives substantial funding because they work at a prestigious university while the other receives nothing because they are independent, the difference is difficult to justify on the basis of contribution itself.

Similarly, consider two people with comparable essential needs. It is inconsistent to treat our own need as urgent while dismissing the other person’s equivalent need merely because it is less visible to us.

This leads to a useful moral test:

Would I accept the same rule if I did not know whether I would benefit from it?

This resembles the philosophical device associated with John Rawls’s theory of “justice as fairness.” Rawls asks which principles people would choose from a position in which they did not know their social class, wealth, abilities, or institutional status. The point is not that Rawls provides the only theory of fairness, but that impartiality requires us to judge rules independently of our expected personal advantage. The Stanford Encyclopedia of Philosophy’s discussion of Rawls explains this approach as a system of equal basic rights and fair social cooperation.

Is Fairness the Same as Equality?

No. Equality is one important component of fairness, but the two concepts are not identical.

Strict equality gives everyone the same amount. Fairness may instead require distributions based on:

• basic needs;
• useful contribution;
• effort or sacrifice;
• responsibility;
• equal opportunity;
• compensation for disadvantage;
• benefits produced for other people.

For example, distributing an identical research budget to every person on Earth would be equal in a narrow mathematical sense, but it would not necessarily be an effective or fair method of financing research. Some people have produced extensive scientific work; others have not chosen to participate in research. Treating these situations identically would ignore a relevant difference.

Conversely, awarding nearly all research funding to people with prestigious institutional affiliations is also unfair. Institutional status may correlate with opportunity, but it is not identical to scientific value.

A defensible funding system therefore needs at least two distinct layers:

1. A social minimum based on human need and dignity.
2. Additional rewards based on useful, verifiable contribution.

AIIM primarily addresses the second layer. It is not intended to replace healthcare, disability support, food assistance, or other need-based social protections. It proposes a fairer method for distributing money intended to reward intellectual and technical contributions.

Moral Fairness: Judging Our Needs and Others’ Needs by the Same Standard

Moral fairness begins with symmetry.

We normally believe that our own suffering matters, our work deserves recognition, and our legitimate interests should not be ignored. But if those judgments are based on characteristics shared by other people—the capacity to suffer, to work, to create, and to pursue meaningful goals—then the same reasoning applies to them.

This does not require ignoring personal relationships or special responsibilities. A parent may have stronger obligations toward their own child than toward a stranger. A government may have specific legal duties toward its citizens. Nevertheless, special obligations do not make outsiders morally irrelevant.

The central error of unfair judgment is often not explicit hostility. It is asymmetric attention:

• We know the reasons for our own failures but see only other people’s results.
• We interpret our own requests as legitimate needs but describe similar requests from others as entitlement.
• We treat our credentials as evidence of merit while treating another person’s lack of credentials as evidence of incompetence.
• We regard our unconventional ideas as original but another person’s unconventional ideas as unserious.
• We expect others to evaluate our work carefully while dismissing theirs through a title, affiliation, nationality, or social category.

A fair system must reduce this asymmetry. It should ask the same questions about every participant and apply the same evidentiary standard.

Legal Fairness: Equality Before the Law and Non-Discrimination

Legal fairness is narrower than morality. Law cannot enforce every moral obligation, and something can be morally unfair without being illegal. Nevertheless, modern legal systems commonly recognize several core principles:

• equality before the law;
• equal protection;
• non-discrimination;
• procedural fairness;
• the right to know and challenge important decisions;
• consistency in the application of rules.

Article 7 of the Universal Declaration of Human Rights states that all people are equal before the law and entitled to equal protection without discrimination. The International Covenant on Civil and Political Rights expresses a related legal principle in its equality and non-discrimination provisions.

However, equality before the law does not imply that every person must receive the same financial outcome. It means that legally relevant decisions should not be determined by prohibited discrimination, arbitrary privilege, or rules selectively applied to favored groups.

In public funding, legal fairness may therefore require:

• published eligibility rules;
• consistent evaluation criteria;
• reasonable accommodation where required;
• records of decisions;
• safeguards against conflicts of interest;
• opportunities for review or appeal;
• protection against unlawful discrimination.

Traditional research funding can satisfy some of these requirements formally while remaining substantively unequal. Two applicants may technically enter the same competition, yet one may have professional grant writers, institutional endorsements, preliminary funding, established citations, and personal access to reviewers. The other may be an unaffiliated researcher whose work does not fit recognized categories.

Formal permission to apply is not the same as genuinely fair access.

Why Existing Money Distribution Is Often Unfair

Modern economies distribute money through several overlapping systems:

• markets;
• employment;
• ownership;
• inheritance;
• taxation;
• welfare programs;
• philanthropy;
• grants;
• political budgets;
• personal networks.

Each system serves a purpose, but none measures human value or social contribution directly.

Market income often measures purchasing demand, scarcity, bargaining power, ownership, and control over distribution channels. It does not necessarily measure the long-term usefulness of an activity. A foundational mathematical result may initially generate little revenue, while a highly marketed but socially marginal product can produce substantial profit.

Grant systems introduce expert judgment, but they also reward proposal writing, institutional reputation, conformity with established research programs, and the ability to predict results before conducting the work. Much of the money is allocated before the contribution exists.

Charitable giving can address genuine needs, but donors usually lack enough information to compare thousands or millions of possible recipients. Donations therefore concentrate around people and organizations that are visible, emotionally compelling, or professionally promoted.

This produces a general information problem:

Fair distribution requires comparing many people using relevant evidence, but no human decision-maker can examine everyone consistently.

Committees solve the problem by narrowing the pool. Markets solve it through prices. Bureaucracies solve it through categories and documentation. Social networks solve it through trust and familiarity. Every shortcut excludes information and introduces systematic bias.

What AIIM Changes

AI Internet-Meritocracy is being developed as a funding system for scientists and free and open-source software developers. Instead of requiring each participant to win a conventional grant competition, AIIM is intended to assess publicly available evidence of contribution and allocate donated funds proportionally.

The model can be described as:

contribution-first, continuous, evidence-based, and auditable funding.

A participant connects sources documenting their work, such as research profiles, publications, software repositories, and other relevant accounts. The system evaluates the work and estimates its relative contribution. Available funds are then divided according to those assessments.

This differs from traditional grants in several ways:

Traditional grant funding	AIIM model
Money is usually awarded before results	Funding follows documented contribution
Applicants compete in limited rounds	Evaluation can be continuous
Grant-writing ability strongly affects success	Published work is the primary evidence
Institutional affiliation often matters	Independent contributors can participate
Decisions are made by small committees	Evaluation is algorithmic and designed for audit
Rejected applicants may receive little explanation	Scoring factors should be inspectable and contestable
Funding categories are defined in advance	The system can evaluate work across a broader contribution graph

The current AIIM implementation remains a developing beta rather than a completed global financial institution. Science DAO’s product-status page identifies unresolved governance and security work. Claims about its fairness must therefore be understood as claims about the proposed architecture and intended mature system—not proof that every present output is already correct.

Why AIIM Could Be Fairer Than Human Committees

The same basic process can apply to everyone

A committee may unconsciously apply different standards to insiders and outsiders. AIIM can be designed to process comparable evidence through a common evaluation pipeline.

That does not eliminate bias automatically. It does, however, make inconsistent treatment easier to detect statistically and technically.

It can examine more evidence

Human reviewers have limited time. They may read a proposal, a short CV, and selected publications. An AI-assisted system can potentially inspect a much wider body of public work, including dependencies, citations, software use, reviews, corrections, and relationships between contributions.

More evidence does not guarantee a better decision, but it reduces reliance on crude proxies such as degrees, university names, or geographic location.

It can recognize indirect contribution

Scientific and technical progress is a dependency network. A widely used application may depend on a small library; that library may depend on a compiler, protocol, theorem, or data format maintained by people who receive little public recognition.

AIIM is intended to consider these relationships rather than rewarding only the most visible final product. Science DAO describes this as evaluating impact, dependencies, and usage when distributing funding.

It can separate contribution from self-promotion

In conventional systems, people must repeatedly explain why their work matters. This favors confident communicators, professional fundraisers, native speakers of dominant languages, and people with institutional support.

A contribution-based engine can shift attention from the persuasiveness of the application to evidence left by the work itself.

It can fund independent researchers

A person’s lack of a degree does not logically prove that their research is worthless. Neither does residence in a poor country, unemployment, disability, an unconventional career, or exclusion from an institution.

The AIIM non-discriminatory grants model is designed to remove degree requirements and institutional permission as preconditions for evaluation.

It can create an audit trail

A fair decision is not merely one that happens to be correct. It should also be possible to inspect how the decision was made.

Blockchain records can document transfers, while published evaluation rules, model versions, data sources, and governance decisions can make the surrounding process more accountable. This does not prove that the evaluation itself is fair, but it makes hidden manipulation more difficult.

Why “AI Decided” Is Not a Sufficient Definition of Fairness

An AI system can reproduce discrimination found in its data. It can favor work written in widely represented languages, confuse popularity with importance, reward strategic manipulation, or assign false confidence to uncertain evaluations.

Therefore, AIIM should not be called fair merely because it uses artificial intelligence.

The OECD’s Principles on Artificial Intelligence connect trustworthy AI with fairness, human rights, transparency, explainability, oversight, and accountability. UNESCO likewise states that AI actors should promote social justice, fairness, and non-discrimination. These standards imply that a high-stakes distribution engine requires governance, not just a powerful model.

A defensible AIIM implementation should include:

• transparent eligibility and evaluation rules;
• disclosure of important data sources;
• explanations understandable to recipients;
• mechanisms for correcting missing or false data;
• an appeal and re-evaluation process;
• regular testing for demographic, geographic, linguistic, and institutional bias;
• protection against fabricated publications, citation rings, and coordinated manipulation;
• human or decentralized oversight for exceptional disputes;
• public model and policy versioning;
• separation between system administrators and unilateral control of funds;
• independent security and fairness audits.

Without these safeguards, AIIM could become merely a more scalable bureaucracy. With them, it can become a fundamentally different institution: a distribution system in which decisions are open to measurement, criticism, and correction.

Need and Merit Must Not Be Confused

The moral question “Who needs money most?” is not identical to the meritocratic question “Whose work contributed most?”

A person can have great needs without having produced scientific work. Another person may have made a major contribution while already being financially secure. A complete economic system must address both cases, but it should not conceal the distinction.

AIIM’s primary allocation criterion is contribution. Its purpose is to reward scientific and open-source work more accurately than markets, grants, or institutional salaries often do.

However, need can still be incorporated in carefully defined ways. Possible models include:

• maintaining a separate need-based fund;
• applying a minimum payment threshold to meaningful contributors;
• adjusting for unequal access to equipment or infrastructure;
• funding accessibility measures;
• allowing donors to choose between contribution-based and need-sensitive pools;
• ensuring that a recipient’s inability to market themselves does not reduce their evaluation.

This separation improves conceptual clarity. A merit engine should not pretend to be a complete welfare state, and a welfare system should not require people to prove exceptional intellectual merit before receiving basic necessities.

Is AIIM the Fairest Money-Distribution Engine Humanity Has Created?

That conclusion would be premature as an empirical statement. AIIM has not yet distributed a substantial portion of global income, operated at full scale, or undergone the independent audits necessary to establish such a record.

A more accurate claim is:

AIIM is a candidate architecture for the fairest large-scale contribution-based money-distribution engine yet proposed, because it combines universal eligibility, common evaluation criteria, continuous evidence-based assessment, dependency-aware attribution, transparent transfers, and contestable governance.

Its strongest innovation is not AI alone, blockchain alone, or meritocracy alone. It is their proposed combination:

1. AI makes broad evaluation computationally possible.
2. Open Internet evidence allows contributors outside established institutions to be examined.
3. Dependency analysis helps recognize foundational and indirect work.
4. Blockchain records can make distributions publicly auditable.
5. Decentralized governance can limit unilateral control.
6. Continuous reevaluation allows mistakes and outdated assessments to be corrected.
7. Uniform participation rules reduce arbitrary exclusion.

No previous institution has reliably applied all these properties to a global pool of scientists and software developers.

AIIM should therefore be presented neither as an infallible judge nor as a finished proof of perfect justice. It is an attempt to construct something societies have previously lacked: a scalable mechanism for applying approximately the same evidentiary standards to every contributor, regardless of institutional rank.

From Personal Fairness to Computational Fairness

The moral principle behind AIIM begins at an individual level:

Judge another person’s comparable need as seriously as you judge your own, and judge another person’s work by the same criteria you would want applied to yours.

At small scale, conscience can sometimes enforce this principle. At global scale, conscience alone is insufficient. Billions of people cannot personally evaluate one another, and small committees cannot impartially understand the entire structure of knowledge production.

That is why fairness must be embedded in institutions and software.

Computational fairness does not replace moral reasoning. It operationalizes moral commitments by specifying:

• what evidence counts;
• how people are compared;
• which distinctions are relevant;
• how decisions are recorded;
• who can challenge them;
• how mistakes are corrected;
• who governs the evaluator.

The deepest test of AIIM will not be whether its developers describe it as fair. It will be whether people who receive less money can inspect the system, understand the reasons, identify errors, and obtain a meaningful review.

A Fairer Economic Principle

Most existing distribution systems ask one of four questions:

• What do you own?
• What can you sell?
• Whom do you know?
• Which institution approves of you?

AIIM proposes another question:

What useful contribution have you made, according to evidence that can be evaluated under rules applicable to everyone?

That is a significant moral and institutional change.

It does not solve every form of distributive justice. It does not determine basic human rights, replace need-based assistance, or eliminate the necessity of democratic law. It cannot guarantee unbiased outcomes merely by using AI.

Nevertheless, a properly governed AIIM system could make the distribution of research and open-source funding substantially less dependent on prestige, geography, credentials, and personal access. It could recognize contributors whom markets cannot pay and institutions do not see.

The goal is not to declare that a machine possesses perfect moral judgment. The goal is to construct a system that is more consistent, more inclusive, more transparent, and more correctable than the alternatives.

That is the practical meaning of fairness: not identical outcomes, but the same relevant standards, applied to ourselves and to others.

World Science DAO is developing AIIM as open infrastructure for this purpose. Learn more about how AI Internet-Meritocracy distributes research funding, examine its current development status, or support the development of auditable scientific funding.