Skip to content
The ColumnCommentary· No. 3244

Is AI Really Destroying Jobs? A Major Study Says Otherwise

For two years now, a dominant narrative has taken hold of public debate: artificial intelligence is supposedly destroying jobs on a massive

Premium reading
MadMax
Key takeaways
  1. For two years now, a dominant narrative has taken hold of public debate: artificial intelligence is supposedly destroying jobs on a massive
  2. Introduction: The Job-Apocalypse Narrative Wobbles
  3. A Massive Study That Shakes Up Certainties
Transparency

Facts, quotes, and cited links remain in the body. Interpretations are framed as analysis or opinion according to the format.

Introduction: The Job-Apocalypse Narrative Wobbles

A Massive Study That Shakes Up Certainties

For two years now, a dominant narrative has taken hold of public debate: artificial intelligence is supposedly destroying jobs on a massive scale, particularly among young graduates and entry-level positions. A new study published on July 1, 2026 by the financial firm Ramp and the Revelio Labsdatabase seriously dents that thesis.

The analysis, led by Ramp's chief economist, Ara Kharazian, examined data from nearly twenty-two thousand American companies, cross-referencing credit card and bill-payment information with employment records. The result is counterintuitive: companies that invest most heavily in AI grew their headcount by ten percent on average over two years.

The Number That Disrupts the Most: Junior Hiring

Even more surprising, entry-level positions, the ones thought most threatened by automation, grew by twelve percent at these same high-AI-adoptioncompanies. That figure directly contradicts the hypothesis that AI would first replace the routine tasks assigned to new employees.

According to the study, companies with low AI adoption intensity recorded no statistically significant change in their headcount, which suggests the observed effect is indeed tied to intensive use of the technology, not to a broader economic trend.

I have to admit this number surprised me as much as it reassured me. We've been told for months that juniors would be the first victims of automation, and here's a serious study suggesting precisely the opposite at the companies betting most heavily on AI.

The Economic Mechanism Behind These Numbers

AI as a Growth Tool, Not a Replacement Tool

The explanation put forward by the study's authors is a scaling effect: AI reportedly lets companies meet growing demand for goods and services at lower cost, which fuels their overall expansion and, in turn, their need for additional workforce, rather than simply replacing existing employees.

Under this logic, artificial intelligence works as a multiplier of productive capacity, increasing demand for workers to handle a larger volume of business, rather than acting as a simple substitute for human labor in existing tasks.

A Reading That Must Stay Nuanced

This interpretation, appealing as it is, shouldn't obscure the fact that overall headcount growth does not rule out deep transformations in the very nature of the positions themselves. Jobs are probably disappearing in certain specific functions, even if the net balance remains positive for the most technologically advanced companies.

It would therefore be premature to conclude that AI has no disruptive effect on specific categories of workers, simply because the aggregate outcome looks favorable for the companies studied in this particular analysis.

Be careful not to turn this study into an absolute argument that everything is fine. A positive net balance can very well hide painful job losses in certain sectors, offset by gains elsewhere. The average always masks difficult individual stories.

The Striking Contrast With Goldman SachsData

Tens of Thousands of Documented Layoffs Elsewhere

This new optimistic picture stands in stark contrast with other analyses published this year, notably from Goldman Sachs and the layoff-tracking firm Challenger, Gray and Christmas, which count tens of thousands of job cuts directly attributed to artificial intelligenceadoption in 2026.

This methodological and empirical divergence deserves attention: how can two serious approaches produce such different conclusions about the same fundamental economic phenomenon.

Methodologies That Measure Different Things

A plausible explanation lies in the difference in method: the Ramp-Revelio study measures the correlation between AI investment intensity and net headcount growth at the company level, while Challenger, Gray and Christmas data tracks layoffs specifically announced as AI-related, often at large companies undergoing major restructuring.

These two datasets are therefore not necessarily contradictory: it's entirely possible that certain large companies are laying off heavily in specific functions while simultaneously hiring in other parts of their business, a reallocation phenomenon rather than net destruction.

Both studies are probably right, each in their own way, and that's exactly what makes this debate so frustrating for the general public looking for a simple answer to a question that, economically speaking, is anything but simple.

What This Means for American Workers

A Source of Measured Hope for New Graduates

For young graduates entering the job market in an already tense economic climate, this study offers a legitimate glimmer of hope: it suggests that the most technologically dynamic companies aren't necessarily closing the door on entry-level positions, contrary to widespread fear.

That doesn't mean the transition will be easy for everyone. The skills required for these new positions are evolving fast, and workers will need to keep adapting continuously to stay employable in an environment transformed by intelligent automation.

The Real Risk: Polarization of the Job Market

The most serious danger may not lie in net job destruction, but in a growing polarization between workers able to leverage AI tools to boost their productivity, and those left marginalized for lack of adequate training or access to new digital skills.

This polarization, if it holds over time, could dig inequalities far deeper than the simple debate over the total number of jobs available in the American economy.

The real fight of the next decade won't be about whether AI destroys jobs in absolute numbers, but about who gets access to the training needed to navigate this transition without being left behind.

The Geopolitical Angle: Why This Matters for the West

An American Economy That Must Stay Dynamic

In the broader context of global technological rivalry, the American economy's ability to integrate artificial intelligence without triggering mass unemployment is a major strategic stake. A successful AI adoption, one that fuels growth rather than sabotaging it socially, strengthens the West's position against rivals like China.

Conversely, if AI adoption were to trigger mass unemployment and widespread social instability in the United States, it would considerably weaken the country's ability to maintain its technological and economic leadership on the world stage.

China Is Watching This Debate Closely

Chinese strategists are following these internal American debates over AI's social impact very closely, seeking to identify potential flaws in the Western model in order to better position their own approach, one that's more centralized and less subject to free-market constraints and democratic debate.

That's one more reason this seemingly technical economic debate carries a geopolitical dimension that cannot be ignored in any comprehensive analysis of today's global tech competition.

Every time I see this debate reduced to a simple matter of employment statistics, I remind myself it's also a test of the Western economic model's ability to prove its superiority against the authoritarian alternative Beijing is offering.

The Methodological Limits Worth Acknowledging

A Sample That Isn't Representative of the Whole Economy

It's worth noting that the Ramp-Revelio study's sample, though vast at twenty-two thousand companies, comes largely from Ramp's own client base, a business-expense management company, which could introduce a selection bias toward companies that are already technologically advanced or growing fast.

This potential bias doesn't necessarily discredit the study's conclusions, but it calls for caution before generalizing these results to the entire American economic fabric, which also includes many small businesses less connected to modern digital tools.

The Time Factor Remains Uncertain

Another important limitation concerns the study's time horizon, which covers only two years of observation. The longer-term effects of massive AI adoption on employment could differ significantly from this initial trend, especially as the technology becomes more capable and less dependent on human supervision.

Economists themselves acknowledge it's still too early to draw definitive conclusions about the long-term structural impact of artificial intelligence on the American job market as a whole.

Distrusting hasty conclusions is a golden rule in economics, and this study, reassuring as it is, shouldn't make us forget that two years of data are never enough to predict a decade of technological transformation.

What the Companies Themselves Are Saying

Mixed Accounts on the Ground

Beyond aggregate statistics, several tech-company executives have publicly confirmed this trend of hiring alongside AI adoption, explaining that automation tools let them handle a larger volume of business without necessarily shrinking their existing headcount.

Other companies, particularly in more traditional sectors like finance or publishing, report a different experience, where AI has genuinely helped cut certain repetitive administrative positions, without triggering the massive layoff waves initially feared.

A Transition That Varies by Sector

This sector-by-sector variation suggests AI's employment impact can't be boiled down to a single uniform national trend. Every industry seems to be living through its own version of this transformation, with paces and consequences that vary considerably depending on the nature of the tasks involved.

This complexity should push policymakers to avoid hasty generalizations, whether alarmist or excessively optimistic, in favor of a fine-grained, sector-by-sector analysis of intelligent automation's real effects on American employment.

The temptation to reduce everything to a single national narrative, catastrophe or miracle, is the classic trap of media debate on AI. Reality, as often, is far more nuanced and deserves resistance to oversimplification.

The Role of Public Policy in This Transition

Investing in Training Rather Than in Fear

If this study confirms that AI can coexist with job growth, that doesn't relieve policymakers of their responsibility to invest massively in vocational training and the reskilling of workers whose specific tasks would genuinely be automated.

Public policy should focus on supporting the transition rather than on defensive measures that would try to artificially slow the adoption of technologies whose economic growth potential appears, at least according to this study, largely positive.

The Danger of Excessive, Rushed Regulation

Faced with still-contradictory data on AI's real impact on employment, excessive and rushed regulation, driven by fear rather than solid data, would risk needlessly slowing a growth dynamic the United States needs to maintain its technological edge over international rivals.

The balance to strike remains delicate: protect the most vulnerable workers without smothering the innovation that, according to this study, appears to generate more opportunities than net job destruction at the most advanced companies.

I remain convinced that the best public policy response to AI isn't prohibition or fear, but massive investment in training. That's the only strategy that lets us capture productivity gains without sacrificing the most vulnerable workers.

The Media Debate and Its Excesses on Both Sides

The Alarmism That Dominated the Public Narrative

It must be acknowledged that the dominant media narrative, since the explosion of generative AI, has leaned heavily toward alarmism, amplifying every automation-related layoff announcement without always contextualizing these figures against overall job growth in tech sectors.

This anxiety-inducing media coverage has probably helped shape a public perception of AI as an existential threat to employment, a perception this new study significantly tempers, without fully invalidating it.

The Opposite Risk of Blind Optimism

Conversely, it would be just as reckless to now swing toward blind optimism that ignores the real, sometimes painful transformations experienced by certain workers in sectors specifically hit by intelligent automation, despite the overall positive picture painted by this study.

The truth, as often in economics, probably sits somewhere between these two narrative extremes, in a complex gray zone that simplified media debates too often struggle to represent faithfully.

Neither doom nor triumphalism: that's the hardest balance to strike in the public debate on AI, and yet it's probably the only one that honestly reflects the real complexity of this ongoing economic transformation.

What Investors Are Taking Away From This

A Positive Signal for Tech Markets

On financial markets, this study was welcomed favorably by many analysts, who see it as further confirmation that massive investments in artificial intelligence by large tech companies are generating concrete economic returns rather than just a speculative bubble disconnected from real value creation.

This optimistic reading could bolster investor confidence in companies that keep announcing considerable spending on AI infrastructure, a sector that has raised growing questions about the sustainability of its current investment pace.

Caution Still Warranted for Markets

Despite this encouraging signal, the most seasoned investors continue to call for caution, noting that a single study, however methodologically rigorous, isn't enough to dispel all the uncertainties surrounding the long-term profitability of the massive investments currently committed to artificial intelligence worldwide.

This caution remains healthy in a context where the valuations of certain tech companies have reached levels that worry a significant portion of the international financial community.

Markets love good news that confirms bets they've already made, but the caution of the most seasoned investors strikes me as the healthiest signal in this sometimes excessive chorus of enthusiasm around AI.

The Dissenting Voices That Shouldn't Be Ignored

Economists Who Remain Skeptical

Several labor economists continue to voice reasoned skepticism toward these optimistic conclusions, pointing out that net headcount growth could mask deep qualitative shifts in the very nature of the jobs created, often less stable or lower-paid than the positions historically replaced.

These dissenting voices deserve to be taken seriously, since they remind us that aggregate statistics, however reassuring they may look, can conceal much more complex and sometimes troubling realities at the level of individual affected workers.

The Importance of Tracking This Trend Over Time

Faced with these conflicting analyses, the only reasonable stance is to keep closely tracking how these indicators evolve in coming years, rather than drawing definitive conclusions from a single study, however methodologically solid it may be.

It's this ongoing analytical vigilance that will eventually allow us to distinguish durable structural trends from mere cyclical fluctuations in this major economic transformation still unfolding.

I always prefer analytical humility over displayed certainty. This study is valuable, but it's only one piece of the puzzle, and anyone claiming to have the definitive answer on AI's employment impact is probably lying, consciously or not.

What This Reveals About Our Collective Relationship With Technological Fear

A Recurring Anxiety Throughout History

Economic history is full of similar episodes where new technologies triggered waves of panic about their employment impact, from agricultural mechanization to office computerization, without those initial fears always translating into the catastrophes predicted at the time by their most alarmist critics.

That doesn't mean every current concern about artificial intelligence is unfounded, but it calls for a certain historical humility given our collective tendency to overestimate the immediate negative consequences of major technological transformations.

The Responsibility Not to Give In to Collective Panic

Commentators, policymakers, and the media share a common responsibility: not to give in to collective panic that could lead to rushed, counterproductive policy decisions, while remaining vigilant enough to protect workers genuinely affected by these deep economic transformations.

Striking this delicate balance between legitimate vigilance and excessive panic remains one of the major challenges of our era in the face of the unprecedented acceleration of artificial intelligence capabilities across practically every economic sector.

History teaches us humility in the face of technological panics, but it also teaches us that certain transitions have genuinely crushed lives. Striking the right balance between these two contradictory lessons remains our most urgent collective responsibility.

Toward a Transformed but Resilient American Economy

A Renewed Potential for Global Leadership

If this study's conclusions hold over time, they would strengthen the United States' position as the global leader in successfully integrating artificial intelligence into a dynamic economy capable of generating growth and employment simultaneously, a model few other nations currently seem able to replicate.

This renewed leadership potential is a major strategic asset in the global tech competition, particularly against rivals like China that are betting on more centralized approaches less subject to traditional market mechanisms.

Vigilance That Must Remain Constant

Nevertheless, this positive trajectory must never become an excuse to relax the vigilance needed in the face of the deep social transformations that the massive adoption of artificial intelligence continues to trigger across practically every sector of the contemporary American economy.

It's this combination of measured confidence and continued vigilance that will, hopefully, allow us to navigate this major technological transformation without repeating the social and economic mistakes of previous industrial revolutions.

I want to believe in this positive trajectory for the American economy, but I refuse to let my guard down. Vigilance and measured optimism aren't contradictory, they are, on the contrary, the two indispensable faces of the same collective responsibility.

What This Study Changes in the Battle of AI Narratives

A Necessary Counterweight to the Dominant Alarmist Discourse

For months, headlines devoted to artificial intelligence have swung between technological fascination and existential dread over the announced disappearance of millions of jobs, a discourse that has taken hold of public space without always being checked against solid empirical data representative of the entire American economic fabric.

The Ramp-Revelio Labs study doesn't claim to close this debate, but it brings a valuable statistical counterweight, grounded in tens of thousands of real companies rather than theoretical projections or isolated anecdotes widely relayed by a press sometimes more drawn to sensationalism than to rigorous economic nuance.

An Invitation to Measure Better Before Concluding

What this research mainly invites us to do is measure the ongoing transformations more carefully before drawing definitive conclusions, whether optimistic or pessimistic, about AI's real impact on American employment as a whole.

Policymakers, unions, and business executives would all benefit from relying on this kind of granular data rather than on intuitions or diffuse fears, legitimate as they may be, to guide their strategic choices in the years ahead.

I distrust the prophets of technological doom just as much as the starry-eyed evangelists of frictionless progress. The truth, as often, demands that we resist the temptation of the simple narrative and accept the uncomfortable complexity of the numbers.
One last word before wrapping up: this study must not become an excuse to ignore the workers genuinely hit by automation. Statistical nuance doesn't erase the very real individual stories of job loss.

Conclusion: A Study to Take Seriously, Without Excessive Enthusiasm

An Encouraging Signal but Not Definitive Proof

The Ramp-Revelio Labs study makes a valuable, rigorous contribution to a debate too often dominated by extreme narratives, whether catastrophist or triumphalist. Its conclusions deserve to be taken seriously, without definitively closing a social and economic debate of such complexity.

Reality, as often, probably lies in an intermediate zone between fears of massive job destruction and the optimism of a perfectly smooth transition toward an economy boosted by artificial intelligence with no significant negative social consequences.

The Importance of Continuing to Observe and Adapt

What matters now is to keep closely tracking how these economic indicators evolve in the months and years ahead, while resolutely investing in the training and reskilling policies that will let American workers fully benefit from this major technological transformation, rather than passively suffering its harshest consequences.

It's this combination of analytical rigor and proactive policy action that will ultimately determine whether this transition toward an AI-transformed economy results in broadly shared prosperity or in new, deep social fractures.

By Maxime Marquette, columnist

Columnist's transparency note

My Acknowledged Biases and My Method

I write this column convinced that the West, and particularly the United States, must keep leading the global tech race against its strategic rivals. This bias leads me to look favorably on studies suggesting a successful integration of artificial intelligence into the American economy, while striving to honestly present the methodological limits and legitimate dissenting voices.

My method relies on data published by Ramp and Revelio Labs, as well as on specialized journalistic coverage of economics and technology. I did not have access to the study's raw data, nor did I conduct direct interviews with its authors.

What I Don't Know

I do not know whether the trends observed over two years will hold over a longer horizon, nor how they will evolve as artificial intelligence capabilities keep advancing rapidly. I also remain uncertain about how fully representative the studied sample is of the entire American economic fabric.

Sources

Primary sources

Los Angeles Times — Want an AI-proof job? New research says you may be safer at companies embracing technology, July 4, 2026

USA Today — Firms heavily invested in AI hired more, not less, July 1, 2026

Secondary sources

Yahoo Finance — The AI jobs debate just got more complicated, July 2026

Newsweek — AI squeeze on jobs figures, July 2026

CNBC — Technology, ongoing coverage of AI and employment

Reuters — Technology, economic analyses on AI

Get the tech columns

AI, platforms, digital power: the next analyses straight to your inbox.

Cite this article

Maxime Marquette (2026). Is AI Really Destroying Jobs? A Major Study Says Otherwise. MadMax. https://mad-max.co/en/article/l-ia-detruit-elle-vraiment-les-emplois-une-etude-majeure-dit-le-contraire

How does this piece make you feel?
MM
Maxime Marquette
Independent columnist

Maxime Marquette writes most of the analyses and columns published on MadMax — geopolitics, technology, and current events, no filler.

The Newsletter

Enjoyed this piece? Get the next one.

One chronicle a week, straight to your inbox. No noise.

Comments

0 / 2000

Be the first to weigh in.

This article was generated with AI assistance, under human supervision.

Commentary3380 words17 min read