What happened

AI adoption is quietly eroding professional expertise across entire industries, according to a new research paper that borrows a 58-year-old economic concept to explain why. Nolan Lovett, a researcher at NATO Special Operations University, published the study in the journal Human Resource Development Review, arguing that companies replacing entry-level jobs with AI are triggering a slow-motion crisis that won't be visible for years.

Lovett calls it the "tragedy of the cognitive commons," a direct nod to ecologist Garrett Hardin's 1968 "tragedy of the commons." Hardin described how herders sharing a pasture each rationally add one more animal, individually profiting while collectively destroying the shared resource. Lovett argues the same logic now applies to professional talent pipelines: when one company automates its entry-level roles with AI, it captures 100 percent of the efficiency gains, while the cost of eroded expertise gets spread across every organization drawing from the same talent pool.

### Two mechanisms driving the erosion

The paper identifies two distinct ways AI adoption disrupts how expertise gets renewed. The first is straightforward: AI systems directly replace entry-level positions, so junior employees never get hired in the first place. The second is subtler — even when entry-level jobs survive, AI-assisted juniors reach productivity levels that used to take years of hands-on experience. The struggle that builds deep domain knowledge simply never happens, because the AI does the hard cognitive work instead of the human.

Why it matters

The consequences compound because oversight of AI depends on the exact expertise that AI use is wearing away. Lovett calls this the "validation tether": spotting subtle, domain-specific errors in polished AI output requires deep knowledge, not surface-level fact-checking. Research already shows that people who routinely trust AI answers lose the habit of questioning them — and the entry-level jobs that once trained workers to challenge authority and verify claims are disappearing.

### The Human Reserve Paradox

Organizations still need a reserve of deeply experienced professionals for crisis management, edge cases, and moments when AI systems fail. But no single company has enough incentive to maintain that reserve on its own — it's a cost with no individual payoff, exactly like Hardin's overgrazed pasture. Lovett calls this the "Human Reserve Paradox," and it means even workers who complete the pipeline may end up with shallower expertise, having spent their careers orchestrating AI output rather than doing independent cognitive work.

Critically, the damage has a long fuse. Today's senior professionals were trained five to twenty years ago, long before generative AI reshaped entry-level work. Lovett estimates that cuts to entry-level roles starting around 2023 may not fully show up in the expertise pipeline until somewhere between 2030 and 2045. Some fields, like software engineering, face sharper exposure than others, since AI coding tools already handle tasks once reserved for junior developers.

Best AI tools saved weekly in our channel — @aigobySubscribe →

How to use it today

For founders, marketers, and creators, the practical takeaway isn't to avoid AI — it's to use it deliberately instead of letting it quietly replace the learning process. Before automating a junior task entirely, ask whether a human still needs to attempt it first and use AI to check their work, rather than the reverse. This preserves the friction that actually builds skill.

Small teams and solo operators can apply the same logic on a smaller scale by using free, low-stakes AI tools to prototype and learn rather than to fully outsource judgment. A resource like [mykreatool.com](https://mykreatool.com) is useful here: experimenting with free AI tools on real tasks lets you see exactly where AI output needs a human review step, which builds the same validation instincts Lovett warns are disappearing at scale. Treat every AI-assisted task as a chance to sharpen your own judgment, not just a shortcut to skip it.

Who benefits

Businesses that keep some form of structured, effortful training in place stand to benefit most as this trend plays out. Companies that resist fully hollowing out entry-level roles will hold a growing competitive advantage: a bench of professionals who can actually catch AI's mistakes when it matters. Independent experts and consultants who maintain deep, hands-on skills also gain leverage, since their judgment becomes scarcer — and more valuable — as junior pipelines thin out industry-wide.

Educators, coaches, and course creators building AI-literacy and critical-review training are also well positioned. As Lovett's research suggests demand for genuine domain oversight will grow even as AI adoption accelerates, programs that teach people how to audit and challenge AI output rather than just prompt it will find a receptive market among employers worried about the validation tether.

Risks

The central risk is invisible in the short term. Because the professionals validating AI output today were trained before AI reshaped entry-level work, organizations may not notice the erosion for a decade or more — by which point the talent pipeline is much harder to rebuild. Lovett's research suggests the effects of 2023-era entry-level cuts may not fully surface until 2030-2045, giving leaders a false sense of security in the interim.

A second risk is uneven exposure. Fields like software engineering face faster disruption because AI coding assistants already substitute for junior-level tasks, while other professions may erode more slowly but just as surely. Companies that assume "our field is different" without examining their own entry-level hiring trends risk discovering the gap only when a crisis demands expertise nobody trained for.

Conclusion

The tragedy of the cognitive commons reframes AI adoption as a collective-action problem, not just a company-by-company efficiency decision. Each firm that automates entry-level work benefits individually, but the shared pool of professional expertise pays the price — and the bill may not come due until 2030 or later. For entrepreneurs and teams building with AI today, the lesson is to protect the learning loop deliberately, using tools to augment human judgment rather than replace the struggle that creates it.