June 23, 2026

The Jobs AI Is Stealing That Nobody Is Talking About

When people talk about AI taking jobs, they picture a factory floor.

Robots welding cars. Machines sorting packages. Conveyor belts replacing calloused hands. This is the story we have been telling ourselves for forty years. Automation is a blue-collar problem. The educated, the credentialed, the white-collar professional is safe.

That story is wrong. It was always wrong. And now the evidence is impossible to ignore.

AI is not coming for the factory. It already came for the factory. Now it is standing quietly in the lobby of every law firm, hospital, creative agency, financial institution, and corporate headquarters in the world. And most of the people inside do not know it is there yet.

The Legal Profession

Law has always been one of the most protected professions on earth. You needed years of expensive education. You needed a license. You needed to understand precedent across thousands of cases, across decades of interpretation.

AI does all of that in seconds.

Paralegals spent years learning to research case law. They found relevant precedents, organised arguments, prepared briefs. Large law firms employed hundreds of them. AI now does this work in a fraction of the time and at a fraction of the cost. Major firms have already cut paralegal headcount. The ones that have not are planning to.

But it is not just paralegals; Contract review, Due diligence, Document discovery, and Compliance checking. These are tasks that kept entire floors of junior lawyers employed for decades. They are now being automated faster than law schools are updating their curricula.

The junior lawyer who was supposed to do five years of document review before making partner has no documents to review. The ladder has been pulled up.

The Creative Industries

Graphic designers were told they were safe. Creativity cannot be automated. Art requires a human soul.

Tell that to the stock photo industry, which collapsed almost overnight when AI image generators arrived. Tell it to the junior designers at marketing agencies who used to spend their days creating social media graphics, advertising banners, and product mockups. Those jobs are gone. Not being threatened. Gone.

Copywriters who built careers writing product descriptions, email campaigns, and website content have watched AI produce their work in seconds. The market rate for this type of writing has collapsed. Rates that once sustained a middle-class income are now a fraction of what they were three years ago.

Video editors, translators, voice-over artists, illustrators. Each of these professions built thriving freelance economies. AI has undercut the pricing in every single one of them. The people still working are working for less. Many have left entirely.

Finance and Accounting

Entry-level finance jobs have always served a clear purpose. You hired analysts to build models, crunch numbers, and produce reports. You hired accountants to review documents, identify discrepancies, and prepare filings. These were excellent, well-paying jobs for smart graduates who had studied hard.

Goldman Sachs built AI tools that do in minutes what junior analysts spent weeks doing. Other major banks followed. They did not announce redundancies. They simply stopped hiring at the same rate. The jobs did not disappear with headlines. They evaporated quietly, one hiring freeze at a time.

Tax preparation, which employed hundreds of thousands of people at companies like H&R Block and countless independent firms, is being rapidly automated. Routine tax returns no longer need a human. The complexity threshold for requiring human expertise keeps rising. Everything below that threshold is gone.

Medicine and Healthcare Administration

Radiology was once considered among the most secure specializations in medicine. Reading scans requires years of training. It requires pattern recognition built from thousands of cases. It requires the kind of expert judgment that takes a decade to develop.

AI now detects certain cancers in medical imaging more accurately than radiologists. This is genuinely remarkable. It is also genuinely destabilizing for thousands of people who spent a decade and enormous sums becoming qualified to do exactly this work.

Below the doctors, the disruption is even more immediate. Medical transcriptionists who converted doctor-patient conversations into clinical records are being replaced by AI transcription tools that do the same work instantly. Medical coders who translated diagnoses into billing codes are being automated out of existence. Healthcare administration, which employs millions, is being stripped back aggressively.

Technology Itself

Here is the cruelest irony in this entire story. The tech industry told everyone else to learn to code. “Automation is coming for your job, so learn to code.” It was the standard reassurance of the last twenty years.

AI writes code.

Junior software engineers are discovering that the entry-level positions they trained for are evaporating. GitHub Copilot and its successors allow a senior engineer to do the work that previously required a team. Startups that once hired ten junior developers now hire two senior ones and let AI do the rest.

QA engineers, Technical writers, Data entry specialists, and Junior data analysts. These roles are disappearing from job boards right now. Not in ten years. Now.

The Pattern Nobody Is Naming

Look at every professions I have described: Law, Creative work, Finance, Medicine, and Technology. In every single case, the pattern is identical.

AI hits the bottom first. The entry-level positions go first. The junior roles go first. The repetitive, well-defined, high-volume tasks go first. And because these are not yet glamorous or senior positions, nobody writes front-page stories about them. There is no mass rally. There is no political movement. There is just a quiet accumulation of people who cannot find work in the field they trained for.

The senior people are still employed. For now. But here is what is about to become clear. The senior people in law, finance, medicine, and technology got there by doing the junior work for years. That junior work built their expertise. It is how mastery is developed. It is the apprenticeship model that every profession has relied on for centuries.

Eliminate the junior work. In ten years you have no senior people either. You have a generation of nominally qualified professionals who never developed real expertise because AI did their learning for them.

Why It Is Happening Quietly

Companies do not announce that AI is replacing their workforce. It is bad for public relations. It triggers regulatory attention. It demoralizes remaining employees.

Instead they use different language. They say they are “not backfilling” positions when people leave. They say they are “restructuring for efficiency.” They say they are “right-sizing” in response to market conditions. The jobs disappear. The language never mentions why.

This matters because it makes the disruption statistically invisible for longer than it should be. Unemployment numbers do not capture people who graduated and could not enter their field. They do not capture freelancers whose rates collapsed. They do not capture the paralegal who retrained as a project manager but earns thirty percent less. The real damage runs deeper than the official numbers show.

The Middle Class Is the Target

This is not an accident. It is the structural logic of the technology.

AI is most effective at tasks that are well-defined, high-volume, and knowledge-based. These are exactly the tasks that built the middle class in the knowledge economy. These are the tasks that millions of people spent years and significant money training for. These are the tasks that were supposed to be immune to automation.

The jobs least threatened by AI are at the extremes. The truly elite, where uniquely human judgment, relationships, and creativity at the highest level matter. And the truly physical, where robots remain expensive, limited, and impractical. The middle, the competent professional doing solid knowledge work, is exactly where AI hits hardest.

We are not talking about the hollowing out of manufacturing this time. We are talking about the hollowing out of the professional class.

What Needs to Be Said Out Loud

There is a version of this story that is told in the boardrooms of technology companies, and it goes like this: AI will create new jobs. New industries will emerge. Humans will move to higher-value work. It happened with the internet. It happened with every previous wave of automation.

Maybe. But the people saying this are not the ones losing their jobs. They are the ones whose companies are doing the automating. They have every incentive to be optimistic and no personal stake in being right.

The people who should be speaking loudly are the ones affected. The paralegal who cannot find a position. The junior designer whose rates halved. The new graduate who cannot get into their field. The copywriter who retrained and retrained and is still earning less than they were five years ago.

These people exist in their millions. They are just not being asked.

The Question for Our Generation

I am not writing this to argue that AI should be stopped or banned. That is neither possible nor desirable. Tools this powerful, this useful, and this economically advantageous will be deployed. The question is not whether. The question is how and for whom.

Right now, the productivity gains from AI automation are flowing almost entirely to the companies deploying it and the shareholders who own them. The workers whose roles are being eliminated receive severance, if they are lucky, and a suggestion to reskill. The communities built around those professional jobs receive nothing.

A society that allows the gains of automation to be entirely privatized while the costs are entirely socialized will not remain stable. It never has. The history of industrial disruption without redistribution is not a history of smooth transitions. It is a history of social fracture.

We are in the early stages of the most significant professional disruption in living memory. The jobs going quietly right now are just the beginning.

The question is whether we are going to have an honest conversation about it before the damage becomes impossible to reverse. Or whether we will keep pretending it is a factory problem while the professional class quietly disappears.

Originally published on Substack. ← Back to all articles