MIT Just Named 7 White-Collar Jobs AI Will Destroy Before 2028

MIT research on AI job elimination has identified a specific set of high-paying, white-collar roles that are already shrinking — not in the future, not in theory, but right now, in 2026, while most professionals still believe their career is safe.

The list is not what the headlines promised.

It is not robots replacing surgeons or AI writing Oscar-winning screenplays overnight.

What MIT actually found is quieter, more targeted, and far more dangerous for people in specific roles — especially those earning between $60,000 and $120,000 a year.

This article breaks it all down in plain language, using real data, real people, and the actual MIT and Stanford findings that most tech influencers have completely misread.

👉 Free download: Start a 1-Person Business With Claude AI — Free Quick-Start Guide

The Two Things Everyone Gets Wrong About AI and Jobs

Before you panic, or before you dismiss the entire conversation as hype, you need to understand one important idea that MIT researchers actually put a name to.

There are two completely different ways that technology kills jobs.

The first is what MIT’s labor automation study calls a crashing wave — a sudden, total disruption that makes an entire category of work obsolete almost overnight.

Think about the invention of the telegraph, the mass production of penicillin, or nuclear fission.

These were moments when a technology arrived so fast and so completely that the jobs attached to the old method simply ceased to exist at any meaningful scale.

A crashing wave leaves no runway for adaptation.

One day the job is there, the next day it is gone, and retraining cannot move fast enough to catch the workers being swept away.

The second type is a rising tide — a gradual, sustained improvement across a broad problem space that unfolds over years, even decades.

When Netflix launched in 1998, it did not kill Blockbuster in a single afternoon.

When the iPhone launched in 2007, Steve Jobs did not walk off the stage and announce that personal computers were dead by Thursday.

Rising tides give employers and employees time to notice, adapt, and reposition.

MIT’s study draws a sharp line between these two concepts because the difference between them determines whether you have time to save your career or not.

The crashing wave is what Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman are describing when they talk about 50% of all entry-level white-collar jobs being wiped out within five years.

That specific forecast from Dario Amodei, made on multiple public platforms throughout 2024 and 2025, paints a picture of unemployment reaching levels last seen during the Great Depression of the 1930s — levels that in 1933 hit nearly 25% of the American workforce.

But MIT research on AI job elimination draws a very different conclusion.

The actual data suggests that what we are living through right now is a rising tide, not a crashing wave — and knowing that changes everything about how you should respond.

👉 Get Access to: Start a 1-Person Business With Claude AI

What the MIT and Stanford Papers Actually Found

Two papers dropped in close proximity — one from MIT and one from Stanford — and together they form the clearest, most data-driven picture we have of what AI is actually doing to employment in 2026.

The MIT study focused on task-level automation — meaning instead of asking “will AI replace lawyers,” it asked “which specific tasks that lawyers do can AI now perform, and at what cost?”

That framing matters enormously, because most jobs are not one single task.

They are a bundle of dozens of tasks, and AI may be able to handle three of them while still being completely useless at the other nine.

The Stanford study, titled the AI Index Report, looked at labor market trends across low-skill and high-skill job categories following the public launch of ChatGPT in late 2022.

What Stanford found in customer service roles was striking.

Since ChatGPT launched, early-career and early mid-career customer service positions have declined in a measurable, consistent trend.

Companies are pulling back from investing in junior-level employees, and while multiple factors are at play — including post-pandemic overhiring corrections — the timing is difficult to separate from the rise of AI-powered customer interaction tools.

What is even more striking is that MIT research on AI job elimination shows the exact same trend emerging in high-skill roles.

Software engineering — one of the most celebrated career paths of the 2010s — is experiencing a severe contraction at the entry level.

Stanford’s data shows that roles for early-career software engineers are down significantly.

Early mid-career positions have flatlined or declined slightly depending on the dataset.

Meanwhile, mid-career and senior-level positions are actually up, which tells a specific and important story.

AI is not erasing entire professions.

It is compressing the early part of the career ladder — the rung that entry-level workers need to climb in order to eventually become the senior employees that companies still want and desperately need.

The Real People Losing Real Jobs Right Now

Data without faces is easy to dismiss, so here are the stories that bring the MIT and Stanford findings to life.

Timothy McKean, an Irish translator who spent years working for the European Union, lost over 70% of his income when EU translation contracts migrated to AI-powered translation tools.

He described the experience in terms that anyone in a technical field should find chilling — the more AI learned, the more obsolete his years of specialized knowledge became.

Translation work for large institutions is, in the language of MIT’s framework, a crashing wave sector.

The task is discrete, repetitive, scalable, and almost perfectly suited to what large language models do well.

Paulo Delgado, a writer who publicly documented his decision to leave professional writing and return to software development, predicted that AI could never touch high-end quality writing.

He was right — but he discovered that most corporations do not care about high-end writing.

They are perfectly content with functional, serviceable content produced at a fraction of the cost.

That distinction is brutal and important.

AI does not need to match the best human writers to eliminate the income of most human writers.

It only needs to be good enough for the buyer, and for a large portion of the market, it already is.

Corporate writing — the kind that fills websites, product descriptions, internal reports, and marketing emails — is effectively a crashing wave category.

Manashi, a 21-year-old computer science graduate from California, graduated in 2024 into a job market that looked nothing like the one that existed three years earlier.

Computer science graduates in 2021 were fielding multiple six-figure offers before they finished their final semester.

Manashi’s only interview offer came from Chipotle.

She is not alone.

Taylor applied to 5,762 tech jobs, received 13 interviews, and received zero full-time offers.

He applied to McDonald’s to cover living expenses and was rejected for lack of experience.

Audrey had her resume — carefully tailored, expertly formatted, emphasizing human creativity and analytical skill — rejected by an automated screening system within three minutes of submitting it.

These are not anecdotes.

They represent a documented, data-confirmed trend in MIT research on AI job elimination — the early rungs of the career ladder in technical fields are disappearing faster than the educational and retraining systems around them can adjust.

👉 Get Access to: The AI Traffic Vault

The Enrollment Drop That MIT’s Own Campus Is Feeling

Here is a detail that rarely makes it into mainstream coverage of the MIT and Stanford papers, and it is one of the most telling signals in the entire debate.

Enrollment in MIT’s computer science programs has dropped by nearly 67% in certain foundational courses.

Fundamentals of Programming — once one of the most popular courses on campus — had 598 enrolled students in the spring semester of 2023.

By recent counts, that number has fallen to 198.

That is not a minor fluctuation.

That is a generational signal from students who are watching the job market in real time and making rational decisions about whether a computer science degree still pays off the way it did five years ago.

MIT research on AI job elimination, when read alongside this enrollment data, paints a coherent picture.

Students are asking a very reasonable question: if AI can already write functional code, debug common errors, and complete entry-level programming tasks at speed — tasks that GitHub Copilot can now handle 56% faster than an unassisted junior developer — then what is the value of spending four years and tens of thousands of dollars to compete for entry-level roles that may not exist by graduation?

That question does not have an easy answer.

What it does have is an urgent implication for anyone currently in or considering a technology career.

The skills that matter most in 2026 are not the same skills that mattered most in 2019.

And if you are building a career, a business, or a side income without understanding that shift, you are working from an outdated map.

👉 Get Access to: The AI Blog Monetization Quickstart Guide

AI Is Making Junior Workers Better — But Harder to Hire

Here is the paradox at the center of MIT research on AI job elimination, and it is genuinely strange once you see it clearly.

AI tools are making entry-level workers measurably more productive — but those same workers are finding it harder than ever to get hired.

A study of GitHub Copilot — the AI-assisted coding tool developed by Microsoft and GitHub — found that tasks were completed 56% faster when developers used the tool, with the largest performance gains concentrated among less experienced programmers.

In other words, a junior developer with Copilot can now produce work that would have previously required someone with two or three more years of experience.

The same pattern appeared in Stanford’s data on AI-powered call center tools.

Customer service assistants using AI support systems showed a 30% improvement in the number of issues resolved per hour, with the biggest gains again coming from novice and less-skilled workers.

AI is democratizing performance.

It is narrowing the gap between the junior employee and the senior one.

But instead of companies saying “great, now we can hire more junior workers and get senior-level output,” many are saying “great, now we need fewer junior workers because our senior staff is more efficient than ever.”

That logic is short-sighted in a way that several economists and workforce researchers have already flagged publicly.

You cannot manufacture senior engineers without first growing junior ones.

A company that stops hiring junior talent today is quietly engineering a senior talent crisis for itself five years from now.

But in the immediate term, that short-term thinking is exactly what is making the job market so brutal for people entering technical and white-collar fields right now.

The Jobs That Are Truly in Danger vs. The Jobs That Are Not

Let us be direct about the categories, because MIT research on AI job elimination does not suggest that all $100K jobs are equally at risk.

High Risk — Crashing Wave Categories:

Translation and localization work for large institutional clients falls squarely in the danger zone.

The EU has already shifted significant volumes of translation work to AI tools, and Timothy McKean’s story is not an isolated case.

Corporate and commercial writing — the kind that fills web pages, product listings, and internal documents — is being automated at scale.

Companies are not debating whether AI can match their best copywriters.

They are replacing their average copywriters and keeping one editor to manage the output.

Entry-level data processing and analysis roles — pulling reports, formatting spreadsheets, summarizing documents — are being absorbed by tools like Microsoft Copilot, Notion AI, and custom GPT workflows faster than HR departments are updating their job descriptions.

Lower Risk — Rising Tide Categories:

Radiology is a fascinating case because Geoffrey Hinton — widely known as the Godfather of AI — publicly stated in 2016 that AI would outperform radiologists within five years.

A decade later, radiology employment is not down.

At Mayo Clinic, the radiology staff has grown by 55% since Hinton made that prediction, now employing over 400 radiologists.

AI has made radiologists faster, more accurate, and capable of handling higher caseloads — but it has not replaced them.

Software engineering at the senior and mid-career level is following a similar trajectory.

The work is becoming AI-assisted, not AI-replaced, for experienced practitioners.

Customer success management, strategic consulting, complex sales roles, and cross-functional project leadership all require the kind of contextual human judgment, relationship management, and organizational navigation that AI tools are genuinely poor at replicating.

Why the Layoffs You Are Seeing Are Not Entirely About AI

This is the part of the MIT research on AI job elimination conversation that most coverage completely misses.

A significant portion of the tech layoffs between 2023 and 2026 had almost nothing to do with AI capability.

Block cut 40% of its workforce.

Coinbase reduced headcount by 14%.

Oracle eliminated 21,000 positions.

When pressed, the CEOs of these companies cited AI adoption as the reason.

But the Federal Reserve Bank of Atlanta, after reviewing employment data across AI-adopting firms, found that AI investments had not materially moved headcount in either direction.

Only 5% of firms in census-level data reported any employment impact from AI at all.

A separate analysis by Ramp — the corporate expense management platform — found that companies that adopted AI tools actually increased their headcount by 10% over the following two years.

The real driver of the 2023 and 2024 tech layoffs was far less dramatic.

Companies massively over-hired during the pandemic boom of 2021 and 2022, when interest rates were near zero, venture capital was flooding the market, and every tech product was experiencing artificial demand inflation.

When interest rates rose sharply through 2023, unprofitable products were shut down, bloated teams were trimmed, and the correction hit hardest on the most recent hires — which, by definition, were the most junior.

MIT research on AI job elimination acknowledges this complexity.

The data does not support a clean “AI did this” narrative.

What it supports is a more uncomfortable truth: AI gave companies a convenient and media-friendly explanation for layoffs that were largely driven by financial over-extension.

The ATM Lesson That Every Worried Professional Needs to Hear

When ATMs began rolling out across US banks in the 1970s, the conventional wisdom was that bank tellers would be wiped out within a generation.

The machines could handle cash withdrawals, deposits, and balance inquiries without ever calling in sick.

What actually happened was the opposite of what everyone predicted.

Bank teller employment did not decrease.

It transformed.

With lower operating costs per branch, banks opened more branches.

With more branches, they needed more staff.

Tellers shifted away from cash transactions and toward relationship management, loan consultations, and customer service work that ATMs were not equipped to handle.

MIT research on AI job elimination points to this exact dynamic as the most likely long-term outcome for many of the roles currently under pressure.

The power loom is another instructive example.

When mechanized power looms arrived in the early 1800s, they reduced the labor required per yard of cloth by over 98%.

Hand loom weavers saw their employment approach near zero over the following decade.

But factory employment soared.

The machine did not eliminate work.

It transformed it, scaled it, and created entirely new categories of labor around the new production reality.

The machine today is Claude.

The machine today is GitHub Copilot and Codex and the growing ecosystem of AI tools that can handle the mechanical, repetitive, clearly-defined tasks that once required a human hour to complete.

That machine will transform work.

It already is.

But if history is any guide — and MIT’s framework suggests it is — the professionals who will struggle most are not those who face AI competition directly, but those who refuse to adapt their skill set to the new reality around them.

👉 Free download: The Claude AI Digital Product Starter Pack — 10 Done-For-You Prompts for Beginners

What You Should Actually Do With This Information

Understanding MIT research on AI job elimination is useful only if it changes how you act.

If you are currently in a crashing wave category — translation, corporate writing, entry-level data processing, basic coding tasks — the window for repositioning is open right now, but it will not stay open indefinitely.

The most practical move is not to run from AI but to get ahead of it.

Learn to use the tools that are reshaping your field so that you become the human layer that AI output requires — the editor, the strategist, the person who knows what good looks like and can direct the machine toward it.

If you are in a rising tide category — senior technical roles, healthcare, education, legal strategy, complex project management — the immediate threat is lower, but the medium-term requirement is the same.

AI literacy is no longer optional.

The professionals who are pulling ahead in every rising tide field are the ones who have figured out how to use AI tools to multiply their output, not the ones who are waiting to see how things shake out.

If you are currently unemployed or between roles in a field being disrupted, the reframe that matters is this: the fact that AI is making junior workers more capable is an opportunity, not just a threat.

A motivated person who learns to use Claude, ChatGPT, GitHub Copilot, or any of the rapidly expanding ecosystem of AI productivity tools has access to capabilities that would have required a team of specialists five years ago.

That is not a consolation prize.

That is a genuine competitive advantage for the person willing to use it.

👉 Get Access to: The Flipboard Traffic Workflow Kit

The Bigger Opportunity Hidden Inside the Disruption

Here is what the doom-and-gloom coverage of MIT research on AI job elimination almost never discusses.

Every major technological disruption in history has created more economic value and more categories of work than it destroyed.

The printing press eliminated the work of hand-copying manuscripts and created publishing, journalism, advertising, and an entire literary economy.

The internet eliminated travel agents, video rental stores, and classified ad departments — and created social media management, SEO consulting, e-commerce, content creation, and thousands of digital product categories that did not exist before.

AI will follow the same arc.

The jobs it eliminates in the short term are real losses for real people, and minimizing that is dishonest.

But the categories of work that AI is already creating — AI workflow consulting, prompt engineering, AI content strategy, digital product creation, AI-assisted coaching, and AI-powered online businesses — are not small or marginal.

They are the fastest-growing earning categories of 2025 and 2026.

The person who waits for their employer to navigate this transition for them is betting on an institution that has repeatedly shown, through the layoff cycles of the last three years, that it will protect its margins before it protects its workforce.

The person who builds a parallel income stream now — using the same AI tools that are reshaping the job market — is positioning themselves on the right side of the disruption.

MIT research on AI job elimination does not end on a note of despair.

It ends on a note of transition.

And transition, for the person who moves early, is almost always an opportunity.

Final Thought: The List Is Real — But So Is the Exit

MIT’s research is not alarmist propaganda, and it is not a comforting dismissal.

It is a precise, data-grounded look at which tasks AI handles well, which jobs are built on those tasks, and which careers are therefore most exposed.

The $100K jobs on that list are real.

Corporate writers, entry-level coders, translators, junior data analysts, and early-stage customer service roles are all experiencing measurable, documented contraction right now.

But the professionals who are thriving inside this disruption share one common trait.

They stopped waiting to see what would happen to their job and started building something that AI cannot easily replicate — a personal brand, a digital product, a niche expertise, or an audience-based business that earns whether they are working or not.

MIT research on AI job elimination confirms the threat.

What it also confirms, for anyone reading carefully enough, is that the threat is not evenly distributed.

It is concentrated among the people who stand still.

👉 Get Access to: Start a 1-Person Business With Claude AI

We strongly recommend that you check out our guide on how to take advantage of AI in today’s passive income economy.