When Scientific Talent Is Wasted: Underemployment, Skills Mismatch, and How AIIM Could Help in 2026

Getting your Trinity Audio player ready...

When a person with substantial scientific or technical expertise has to spend most of their working life doing a job that does not use that expertise, we usually call it “life circumstances.”

That description is often too passive.

There is another way to look at the same situation: society has invested in, produced, or encountered a scarce capability—and then failed to use it.

I propose a deliberately provocative term for this failure: a “crime of the public.”

I do not mean crime in the legal sense. No individual citizen has necessarily broken a law, and the person’s employer may have done nothing wrong. I mean a collective moral and institutional failure: valuable human ability exists, society could benefit from it, yet our funding and employment systems leave much of that ability unused.

Economists and labour institutions use less provocative terms for related phenomena: skills mismatch, overqualification, skills underutilization, and underemployment. The International Labour Organization defines skills mismatch as a discrepancy between the skills people possess and the skills demanded by their jobs or by the labour market.

The important point is that this is not merely unfortunate for the worker. It can also be a loss for everyone else.

The Problem Is Not Factory Work

Suppose a mathematician, physicist, programmer, or other highly specialized expert works in a factory, warehouse, cleaning job, taxi, or another occupation unrelated to that expertise.

There is nothing inherently wrong or undignified about factory work, cleaning, driving, or manual labour. These jobs are necessary, and the people performing them deserve respect.

The problem is mismatch.

If a highly trained machinist is forced to work as an inexperienced mathematician, that would also be an inefficient allocation of talent. The point is not that one occupation is socially “higher” than another. The point is that scarce capabilities should, when possible, be used where they generate the greatest value.

The International Labour Organization notes that overqualified workers may be unable to use their full productive potential and that skills mismatch can reduce productivity and competitiveness.

OECD research reaches a similar conclusion. Cross-country analysis has found an association between higher skills and qualification mismatch and lower labour productivity, in part because human resources are allocated less efficiently.

More recent OECD research based on the 2023 Survey of Adult Skills likewise reports that productivity depends not only on the skills available in a country but also on how skilled workers are allocated across jobs and firms.

So the problem can be stated economically:

Unused expertise is not only a private misfortune. It can be wasted human capital.

Why “Life Circumstances” Can Hide Institutional Failure

The phrase life circumstances shifts attention toward the individual.

Perhaps the researcher did not obtain the right degree. Perhaps they did not receive a grant. Perhaps they live in the wrong country. Perhaps they cannot move to a major university city. Perhaps their work belongs to an unfashionable field. Perhaps they lack the professional network required to obtain an academic position.

All of these may indeed be circumstances.

But calling them circumstances does not tell us whether the surrounding institutions are well designed.

Consider two statements:

“A researcher could not obtain funding, so they had to take unrelated employment.”

and:

“Our funding system failed to allocate resources to a researcher whose continued research would have produced more social value than the alternative use of those resources.”

The first statement describes an outcome.

The second asks whether the outcome was efficient or justified.

That distinction matters.

Scientific Labour Is Especially Vulnerable to This Problem

Scientific expertise presents an unusual allocation problem because its value can be difficult to recognize in advance.

A conventional employer usually knows what an employee is being hired to produce. Scientific research is different. Important results can be:

• uncertain;
• difficult for non-specialists to evaluate;
• valuable only many years later;
• dependent on obscure earlier work;
• produced outside prestigious institutions;
• valuable to humanity while generating little private revenue for the researcher.

This helps explain why scientific labour cannot always be allocated efficiently through an ordinary employment market.

A researcher may create something with substantial public value while having no straightforward customer willing to pay for it.

This is particularly important in basic research, open-source scientific infrastructure, mathematics, theoretical work, and other domains where the social beneficiary may be humanity as a whole rather than a specific purchaser.

The market therefore faces a difficult question:

Who should pay a person today for knowledge whose beneficiaries may be millions of unknown people in the future?

Traditional research institutions answer this question through universities, grants, research institutes, foundations, and peer-review committees.

These mechanisms perform essential functions. But they also create filters.

If the filters fail, productive scientific talent can remain unfunded even when the work itself is valuable.

From Personal Misfortune to Public Loss

Imagine that a researcher could perform four productive hours of scientific work every day but instead must spend those hours earning a subsistence income through unrelated employment.

It would be incorrect to assume automatically that the researcher’s science is more valuable than the alternative work. That has to be demonstrated.

But if the scientific work really does have greater expected social value, the resulting loss is not confined to the researcher.

Other researchers may lose a theorem, dataset, library, experiment, algorithm, proof, review, or software component that would have accelerated their own work.

Future technologies may arrive later.

A useful research dependency may never be created.

The cost can propagate through a scientific dependency network.

This is why the allocation of research funding is not merely an issue of compassion toward scientists. It is an issue of resource allocation.

Why I Call It a “Crime of the Public”

The phrase is intentionally stronger than skills mismatch.

Again, I am not alleging a legal crime.

Nor am I claiming that every taxpayer or every person who does not donate to science is personally guilty of another person’s career situation.

The concept is closer to a collective moral failure.

If a society possesses mechanisms capable of identifying valuable work, possesses sufficient resources to support some of that work, and nevertheless systematically allocates those resources according to factors poorly correlated with actual contribution, then “bad luck” becomes an incomplete explanation.

There is an institutional choice involved.

The responsibility is distributed among funding structures, governments, universities, donors, markets, evaluators, and ultimately the public systems that sustain them.

“Crime of the public” is therefore a philosophical label for a particular kind of failure:

A society wastes human capability when it allows demonstrably valuable expertise to remain unused for lack of an adequate mechanism to recognize and fund it.

Whether one accepts the word crime or prefers institutional failure, the underlying resource-allocation problem remains.

How AIIM Tries to Address the Problem

The proposed AI Internet-Meritocracy (AIIM) takes a different approach to research funding.

Instead of making a conventional academic position the primary gateway to scientific income, AIIM is designed to evaluate researchers and open-source developers on the basis of evidence about their actual work and contributions.

Science DAO describes AIIM as an experimental funding system intended to distribute money directly to researchers and open-source developers while reducing reliance on traditional funding bureaucracy.

See the AI Internet-Meritocracy overview and the more detailed discussion of AI-based research funding.

The fundamental idea is simple:

The ability to continue valuable scientific work should depend more on demonstrated contribution and less on successfully obtaining a conventional academic job.

If such a mechanism works, the mathematician working outside mathematics, the programmer maintaining important scientific software without adequate compensation, or the independent researcher excluded from normal grant channels could potentially receive resources because of the value of their work itself.

This changes the question from:

“Who employs this person?”

to:

“What has this person contributed, how valuable is that contribution, and what resources would allow them to continue contributing?”

That is a major conceptual difference.

Why Continuous Evaluation Matters

Traditional grants are often discrete.

A researcher applies, waits, receives either a positive or negative decision, works for a defined period, and then applies again.

Scientific contribution, however, is continuous.

Someone may produce an unexpectedly important result between grant cycles. Another person may create a small software library that becomes a dependency of dozens of projects. A researcher outside academia may begin producing valuable work without ever having passed through conventional institutional selection.

AIIM attempts to make funding more closely follow demonstrated contribution rather than a small number of institutional checkpoints.

Science DAO also proposes dependency-aware evaluation: scientific and software contributions may matter partly because later work depends on them. This is relevant to the problem of wasted expertise because a contribution that appears obscure in isolation may be highly important within a wider dependency graph.

AIIM Cannot Simply Assume That Every Expert Deserves Funding

There is an important limitation.

Being intelligent, educated, highly qualified, or self-identifying as a researcher does not automatically justify receiving public or charitable money.

A mechanism intended to prevent talent waste must distinguish between:

1. possessing expertise;
2. performing research;
3. producing valuable work;
4. causing or enabling scientific progress;
5. merely claiming that one could do valuable work if funded.

These are different things.

A system that simply pays everybody who says they are underemployed would create enormous incentive problems.

AIIM therefore faces a harder challenge: evaluating evidence of contribution while resisting prestige bias, gaming, fabricated achievements, manipulation of AI systems, and other failure modes.

That problem is not solved merely by replacing committees with artificial intelligence.

AIIM Is an Experiment, Not a Proven Solution

This distinction is essential.

AIIM has not yet demonstrated that it can optimally allocate scientific salaries at scale.

Science DAO explicitly describes its current evaluations as heuristic and experimental rather than validated measurements of causal scientific or economic impact. The system still requires testing, governance, appeals, auditing, and human oversight.

Therefore the proper claim is not:

“AIIM solves scientific underemployment.”

It is:

“AIIM proposes a mechanism that could reduce scientific underemployment if contribution-based evaluation proves sufficiently accurate, fair, manipulation-resistant, and economically effective.”

That proposition can be tested.

And it should be tested.

The Goal Is Not to Guarantee Scientists Comfortable Lives

There is another important distinction.

Research funding should not exist primarily because researchers deserve pleasant jobs.

The purpose is to enable valuable research.

Those objectives often align: giving a productive researcher sufficient income may allow that person to spend more time doing research.

But the causal direction matters.

AIIM should not fund someone simply because unrelated employment is unpleasant. It should fund them when supporting their scientific or technical work is a good use of available resources.

This makes the argument stronger rather than weaker.

The central claim becomes an economic one:

If society can identify scientific work whose expected value exceeds the cost of enabling it, failing to fund that work is an allocative failure.

A Better Allocation of Human Capability

Societies devote enormous effort to producing expertise.

People spend years learning mathematics, physics, biology, computer science, engineering, medicine, and other difficult disciplines.

Universities, families, governments, charities, and individuals invest resources in that training.

Yet training people is only half of the problem.

Their capabilities must also be used.

ILO research treats skills mismatch partly as a problem of underutilized human capital, while OECD research indicates that the allocation of skilled workers is related to productivity.

This suggests a broader principle:

A society is not using its scientific resources efficiently merely because it has educated people. It must also create mechanisms that allow valuable expertise to become productive work.

AIIM is one attempt to create such a mechanism.

It may fail. It may require substantial modification. Human oversight may remain indispensable. Some traditional institutions may outperform it in particular contexts.

But the underlying problem exists independently of whether AIIM succeeds.

From “Life Circumstances” to a Scientific Funding Problem

When an expert spends years unable to use valuable expertise, it is tempting to treat the situation entirely as private biography.

Sometimes that is correct.

Sometimes the person’s preferred research is simply not valuable enough to justify funding.

Sometimes there are more productive uses for scarce resources.

But sometimes the problem is institutional.

A person capable of producing socially valuable knowledge can be prevented from doing so because our mechanisms for recognizing, employing, and funding that person are inadequate.

At that point, “life circumstances” becomes an evasive description.

We should at least ask the harder question:

Was valuable human capability wasted because society lacked a sufficiently good mechanism for recognizing it?

If the answer is yes, then we should not merely sympathize with the individual.

We should improve the mechanism.

That is the problem that AI Internet-Meritocracy is intended to address.

Readers interested in the proposed mechanism can explore the AI Internet-Meritocracy model, compare it with traditional research-funding approaches, or learn how to support independent science through Science DAO.

External sources

The concept discussed here is related to established research on skills mismatch and human-capital allocation:

• International Labour Organization: What is skills mismatch and why should we care?
• International Labour Organization: Skills mismatches
• OECD: Labour Market Mismatch and Labour Productivity
• OECD: Adult Skills and Productivity — New Evidence from PIAAC 2023