Who Benefits Most from AI: Highly-Skilled or Less-Skilled Workers?

In early language model studies, generative AI seemed to produce  larger  gains for less-skilled or less-experienced workers.

Later studies now show different outcomes. Apparently, studies now also suggest that workers with greater competence using AI seem able to capture disproportionately large benefits.

That seems quite contradictory. One potential answer might be that two different processes are at work. 

The first-generation effect of LLMs may be basic skill compression. The least-competent workers become more productive as AI lifts basic performance.

A second-generation effect might involve a new level of skill polarization around AI competence. As often happens, the most-productive workers then become even more productive, even as basic skill competence of the least-skilled also rises. 

The process might be akin to the “strength begets strength” phenomenon we often see with firms in general, which is that the best-managed firms also tend to be the organizations able to reap the most benefit from any new technology. 

Study	Setting / sample	Finding about skill differences	Implication
Noy & Zhang (2023), Science	453 college-educated professionals doing writing tasks	ChatGPT cut time by 40% and increased quality by 18%. The gains were largest among weaker writers, reducing inequality between workers. (DOI)	Strong evidence against the idea that higher-skilled workers benefit more—at least for standardized writing.
Brynjolfsson, Li & Raymond (2025), Quarterly Journal of Economics	5,172 customer-support agents	Productivity increased about 15%. Less-skilled and less-experienced workers gained substantially more; the lowest-skilled group gained roughly 30–35%, while the highest-skilled saw little benefit and sometimes small quality declines. (OUP Academic)	AI can transfer the practices of high performers to low performers, effectively compressing skill differences.
Dell'Acqua et al., BCG/Harvard "Jagged Frontier" experiment	758 consultants performing realistic consulting tasks	AI increased performance substantially on tasks within its capability frontier, with particularly large benefits for consultants who initially performed less well. But AI could also hurt performance on tasks outside its frontier.	AI can be an equalizer, but only if workers know where the technology is reliable.
Humlum & Vestergaard (2024), Denmark	Survey of 100,000 workers in 11 occupations	ChatGPT adoption was higher among younger, less-experienced and higher-achieving workers, particularly men. (IZA)	Here the inequality is partly an adoption/selection effect: workers with greater capability may be more likely to use AI.
Cruces et al. (2026), NBER	Randomized experiment, 1,174 adults, workplace-style problem solving	AI improved everybody's performance, but lower-education participants gained substantially more. The initial performance gap of 0.548 SD fell to 0.139 SD with AI—roughly three-quarters of the gap disappeared. Higher-education participants nevertheless used AI somewhat more effectively. (National Bureau of Economic Research)	Very strong recent evidence that GenAI can reduce education-based productivity inequality, while leaving an important advantage in AI utilization to better educated users.
Idan & Anand (2026)	Randomized experiment with early-career knowledge-worker analogs learning a technical domain	Average performance rose with an LLM, but gains were highly uneven. GPA and prior knowledge did not predict gains; AI Interaction Competence (AIC) did. High-AIC users obtained outsized gains while low-AIC users sometimes received little or negative marginal benefit. (SSRN)	This is probably the strongest evidence for the proposition in your question—but it changes "skill" into skill at working with AI.
Freund & Mann (2026)	Model of occupational/task-specific skills and GenAI exposure	Moderate AI exposure can benefit workers, while high exposure can hurt; effects vary greatly within occupations. Their model projects larger gains for lower earners and higher returns to social skills. (IZA)	Again, the aggregate prediction is not simply "AI favors the highly skilled." It depends on which skills remain complementary to AI.
Bloom et al. (2024), NBER/IZA	Theoretical model of AI and skill premia	Unlike empirical workplace experiments, the model assumes AI is particularly useful for high-skill tasks. Depending on substitution relationships, AI can nevertheless reduce the skill premium. (National Bureau of Economic Research)	Theoretical models don't necessarily predict widening inequality even when AI works disproportionately on high-skill tasks.
Humlum & Vestergaard (2025/26), Denmark	Administrative labor-market records linked to ChatGPT adoption	Despite widespread adoption and reported productivity improvements, they find no detectable earnings effect, with effects larger than 2% ruled out two years after ChatGPT's launch. AI is changing tasks more than aggregate earnings so far. (National Bureau of Economic Research)	Productivity gains do not automatically translate into higher wages for high-skilled workers—or anybody else.

The research arguably makes more sense  if "skill" is divided into three different things:

• Underlying job skill: how good you were at the task before AI

• Domain knowledge: how well you understand the subject being worked on

• AI-use skill: how well you can prompt, interrogate, evaluate, correct and integrate the model's output.

The first category has generally not favored high-skilled workers in the experiments, but the third category increasingly appears to do so.

So it seems AI can reduce the productivity gap between workers while simultaneously creating a new gap between good and bad AI users.

The 2026 Idan-Anand experiment is especially interesting because it explicitly finds that prior knowledge and GPA were not good predictors of who gained most from the LLM. 

What mattered was AI interaction competence; the ability to elicit, filter and verify the model's output.

That might be quite a different matter than the ability of a new employee to master basic skills.

Consider customer support roles:

• Top worker without AI: 100 units

• Bottom worker without AI: 50 units

• AI increases bottom worker to 70

• AI increases top worker to 105

In such instances, we can argue that productivity inequality is lessened. The least-skilled workers perform better. But the performance of the best workers also increases. 

But such instances also are of job categories where the “best” level of performance might be effectively limited. 

Other job roles involving higher cognitive input and output do not have the same degree of limits on “best” performance. In other words, there is no obvious ceiling on output quality. 

In those instances, AI is not going to help a person with cognitive, creative or other limits to approach the possible performance of workers with more innate capabilities. 

The upside from AI might  depend less on whether a job is "high skill" than on whether the job has a high ceiling on the quality of its output.

The key distinction is between tasks with an externally imposed performance ceiling and tasks where better reasoning, creativity, judgment, or synthesis can keep improving the result.

Consider a customer-service representative processing a standard billing inquiry.

There may be a fairly objective definition of a good answer:

• correct information;

• appropriate procedure;

• polite interaction;

• resolution of the customer's problem.

Once the AI gets the worker from, say, 80 percent to 95 percent effectiveness, there isn't necessarily much additional value in getting to 98 percent.

The value of being 10 times better than average is limited because the task itself doesn't have much room for exceptional performance.

Non-routine knowledge work arguably is quite different. Consider a:

• scientist developing a new drug;

• engineer designing a new semiconductor;

• entrepreneur developing a new business;

• investment analyst finding an overlooked opportunity;

• architect designing a building;

• software engineer designing a new system;

• lawyer developing an unusual legal strategy;

• management consultant developing a corporate strategy;

• researcher developing a new theory.

There isn't necessarily a point at which the answer becomes simply "correct” or “good.”

Exceptional or revolutionary value creation can happen, with enormous implications for value.

These are jobs where:

• the problem isn't completely specified

• there are many possible approaches

• success requires combining information from multiple domains

• there isn't an obvious algorithm for solving the problem

1. better answers can have enormous economic value.

In those scenarios, where upside is unlimited or bounded, AI might make a huge difference.

For jobs where a human can produce 0 to 100 units of value, and AI raises the average worker from 50 to 75, that’s helpful.

But in other instances, where a human can produce 0 to 100 units of value today, but potentially 1,000 or 10,000 if they discover something exceptional, the order of magnitude upside is where AI can produce exponentially-better outcomes.

Creative work has a peculiar economic property: “quality” is unrestricted or uneven. 

• A brilliant advertisement can transform a brand

• An exceptional bit of software can create an entirely new category

• An exceptional financial bet can produce enormous returns

• A breakthrough scientific paper can create an entirely new field. 

“Adequate” is one thing. “Transformative” is something else.

The conventional automation story is that machines take the routine jobs while humans retain the non-routine jobs.

But the generative-AI story might also be that AI takes the routine portions of non-routine jobs, leaving humans with increasingly open-ended problems:

• A consultant focuses on business options

• A programmer asks “What should we build?”

• A scientist asks “What should we investigate next?”

• A CEO continues to make decisions about what the company makes

• An entrepreneur decides which possibilities are worth pursuing.

So whether AI helps lower-skilled or highly-skilled workers more, the greatest long-term economic upside of AI might occur where AI can dramatically expand the number and quality of solutions that a capable human can explore, while the value of an exceptional solution has no obvious upper bound.

So open-ended, non-routine work may ultimately be a much more important AI opportunity than routine automation.

Movie Producers are Risk Averse, So Revenue-Leading Films are Franchises and Sequels

An inflation-adjusted list of the top-25 domestic U.S. movie releases suggests a couple of things about what producers think will succeed in the market. 

• Every single film is a sequel, prequel, or adaptation of pre-established content

• Pre-sold recognition matters. Every title arrives with built-in awareness; a built-in audience that doesn't need to be persuaded a movie is worth seeing, only reminded it exists.

• Front-loading of demand: The shift from Titanic's 2,674-theater platform release in 1997 to Spider-Man: Brand New Day's 4,487-location release is intentionally designed  to convert cultural anticipation into a single concentrated weekend before word-of-mouth or piracy or a competing release can erode it.

• The "theatrical-only event" thesis. Producers are betting on opening night as a cultural event

• Superhero and family-adaptation franchises such as Star Wars or animation are specifically designed to appeal broadly.

source: Visual Capitalist 

Of course, these successes are also pre-determined, to an extent.

Studios only produce and market films when they already believe the content has appeal (it has sold in the past), not that it independently reveals what audiences at large "want." 

Other genres or titles are not given the marketing spend or theater count to compete in the first place. 

The list measures where the industry chose to concentrate risk capital.

OpenAI Pauses Astra Release

OpenAI’s pause of releasing Astra (ChatGPT 6) might be significant because it suggests frontier language models may be approaching a level of autonomous action with potential cybersecurity implications that require further safeguards. 

OpenAI hasn’t suggested the model already is at a “critical” point; only that it might be. 

OpenAI defines “critical” cyber capability as being able to autonomously develop functional zero-days across many hardened critical systems, or devise and execute novel end-to-end attacks against hardened targets from only a high-level goal. 

So OpenAI is applying a development-stage safeguard, not adding a deployment disclaimer. 

The point is that OpenAI, speaking about Astra, “cannot rule out” autonomous exploits.  

The broader takeaway is whether developers can reliably constrain frontier models under adversarial conditions.

Virtually Nobody Believes in Completely-Unfettered Free Markets: the Issue is When to Intervene

There are lots of reasons why private equity investment in healthcare, child care, nursing homes or veterinary medicine is not much different than private equity in any other industry. PE normally looks for investment opportunities in industries and firms that are fragmented, mismanaged in some way, with room to grow and often featuring steady cash flow. 

The question is whether the financial structure and investment horizon of private equity ownership are well suited to organizations whose primary outputs include public goods such as health, safety, education, or care for vulnerable populations.

This seems to be another instance where we might encounter the idea of free markets needing a bit of management.

In such firms and industries, the concern is that firms may reduce staff, lower safety standards, or limit care to boost profit. 

Also, prices often rise for patients, students, or tenants after a buyout.

In other cases, the concern is a loss of local focus, as decisions move from local leaders to distant corporate offices.

Nursing homes and low-income clinics often face high risks of neglect because heavy debt loads push cash flow to debt service rather than customer care. 

So hospitals and nursing homes see debates over staffing levels and patient outcomes.

In housing markets rent increases and tenant displacement are issues. In education or child care, issues often include firms prioritizing fees over learning tools.

One study spanning eight countries, but 85 percent focused on the United States. Of the 55 cases, nursing homes were the most commonly studied healthcare setting. The analysis included:

• Nursing homes (17)

• Hospitals (9)

• dermatology (9)

• ophthalmology (7)

• multiple specialties or general physician groups (5)

• urology (4)

• gastroenterology (3)

• orthopedics (3)

• surgical centers (2)

• Fertility (2)

• obstetrics and gynecology (2)

• Anesthesia (1)

• hospice care (1)

• oral or maxillofacial surgery (1)

• Otolaryngology (1)

• plastics (1). 

As you might expect, given PE outcomes in other industries, “PE ownership was most consistently associated with increases in costs to patients or payers,” the study found.

“Additionally, PE ownership was associated with mixed to harmful impacts on quality,” the report says. In some instances, PE ownership was associated with reduced nurse staffing levels or a shift towards lower nursing skill mix.

“Health outcomes showed both beneficial and harmful results, as did costs to operators, but the volume of studies for these outcomes was too low for conclusive interpretation,” the authors conclude.

source: Burch et al 

“No consistently beneficial impacts of PE ownership were identified,” the study suggests. 

That is not to argue for barring PE involvement, but many observers might agree some limits might be desirable.

Private equity typically seeks to create value over a relatively short investment horizon (often three to seven years). 

The concern might be whether some of these financial tools create incentives that conflict with long-term service quality in some industries perceived to have social value with a public character. 

The PE playbook is fairly clear: restructure a business to create higher marketplace value. 

PE practice	Potential business benefit	Possible social concern
Reduce labor costs	Higher margins	Lower staffing ratios
Replace senior staff	Lower payroll	Loss of experience
Centralize purchasing	Lower costs	Lower flexibility or quality
Increase debt	Higher investor returns	Less financial resilience
Sale-leaseback of real estate	Unlock capital	Higher fixed operating costs
Roll-up acquisitions	Economies of scale	Reduced local competition
Aggressive billing	Higher revenue	Higher costs for patients or insurers
Short holding period	Faster capital recycling	Less investment in long-term quality

The key issue is that many quality investments—training, staffing, preventive maintenance, or workforce retention—generate returns over many years, while PE investors often realize returns much sooner.

Among all sectors, nursing homes have been studied most extensively, and there is some evidence of less-desirable outcomes. 

Multiple peer-reviewed studies have found associations between PE ownership and:

• higher hospitalization rates

• higher emergency department use

• increased deficiencies cited by regulators

• reduced staffing levels or changes in staffing mix

• higher mortality in some studies.

At the same time, some studies found little change in certain clinical processes, and a minority found no measurable decline in quality. Overall, recent systematic reviews conclude that the balance of evidence points toward mixed but generally less favorable quality outcomes after PE acquisition. (BMJ)

This does not mean every PE-owned nursing home performs poorly. Rather, ownership structure appears to influence average outcomes across large samples.

Evidence remains mixed across specialties, but systematic reviews generally find that PE ownership is often associated with higher costs, while quality effects vary by sector and study.