Are AI Layoffs Real — Or Just a Convenient Excuse?

Mass unemployment hasn’t arrived. But the claim that “nothing has changed” no longer holds up either.

Five years ago, most people answered the question “should I go into tech?” the same way: yes. Learn Python, build a few projects, land your first job, gain experience, grow. It wasn’t a guarantee, but it worked well enough that millions of people built their plans around it. Today the answer isn’t so obvious — not because developers are no longer needed, but because nobody can honestly say what that first career step will look like in three or four years.

That’s why conversations about AI-driven layoffs are so charged. Some say artificial intelligence is already taking jobs. Others point out that mass unemployment isn’t showing up in the data. Both sides have a point. And both are wrong.

AI-driven job losses rarely look like a scene from a science fiction film, where a robot walks in and a human packs up their desk. More often it’s quieter: a vacancy doesn’t open up, a contractor’s agreement isn’t renewed, a team doesn’t get expanded, a junior role gets folded into a task for a mid-level specialist working with Claude or Copilot. For the statistics, almost nothing happened. For the person involved, everything may have changed.

So are people actually being laid off?

The short answer is yes. Klarna claimed its AI assistant was doing work equivalent to hundreds of customer support employees. IBM openly discussed automating parts of its HR and back-office functions. Duolingo shifted to an AI-first approach and cut a portion of its external contractors. Chegg lost a significant share of its business after students started getting answers directly from ChatGPT. This isn’t speculation about the future — it has already happened.

But it’s important not to oversimplify the picture. At Chegg, AI didn’t “replace employees” one-for-one — it attacked the business model itself. If a student can get an answer from ChatGPT, why pay for a homework-help subscription? At Duolingo, the clearest signal was about contractors, not full-time staff — they’re typically the first to go, because their agreements can simply not be renewed without much drama. At Klarna, the key mechanism was different: AI reduces the need for new hiring through attrition and non-backfilling. People leave, and the company doesn’t open new positions. That’s also a reduction in headcount. It just looks a lot less dramatic from the outside.

So why hasn’t unemployment spiked?

Nearly every week brings a new AI tool, and companies keep talking about productivity and efficiency gains. Yet we don’t see millions of people out of work specifically because of AI. Labor markets haven’t collapsed. There are no queues at unemployment offices. Why?

Because AI is reshaping the job market far more quietly than the headlines suggest. Imagine a company with a hundred employees. A year ago, when someone left, the company opened a new position. Today, the manager might decide to try getting by without a replacement. If it works — the vacancy simply disappears. No press release, no dramatic announcement, no official layoff notice. For the statistics, nothing happened. For the labor market, quite a lot did.

It’s not just jobs disappearing — it’s the rungs on the ladder

Companies aren’t firing three junior developers. They’re just not hiring them anymore. They’re not letting the copywriter go — they’re stopping some of their freelance content orders. They’re not shutting down the support team — they’re letting AI handle part of the ticket volume and not opening new positions. On paper, everything looks fine: efficiency is up, costs are down, investors are happy. But somewhere nearby, the first rung of the career ladder is disappearing — the very rung that nearly every senior developer, manager, and expert once started from.

Every programmer once fixed small bugs. Every marketer once wrote simple copy. Every analyst once built basic spreadsheets. Every lawyer once worked through routine documents. If those entry-level tasks disappear, an uncomfortable question follows: where do tomorrow’s experts gain their experience? Nobody has a convincing answer yet.

Who feels it first

The greatest pressure is visible in roles built around repetitive intellectual tasks: translators, copywriters, content teams, customer support, administrative roles, junior developers, and junior analysts. This doesn’t mean these professions will vanish — but the market is already less willing to pay for baseline work at the same rates as before. Writing is still needed, but “write another templated SEO article” no longer carries the value it once did. Programming is still needed, but boilerplate code, documentation, tests, and small bug fixes no longer look like a reliable training ground for newcomers. AI doesn’t destroy a profession in a single blow — it undercuts the cheap, repetitive, and educational tasks, and then the entire profession is forced to reorganize around what’s left.

But maybe the problem isn’t really AI?

The tech industry started cutting headcount before ChatGPT became a mainstream tool. Many companies over-hired during the pandemic, cheap capital dried up, and investors started demanding profitability again. Some of these layoffs would have happened regardless of AI — and that’s true. But there’s an important distinction: in previous downturns, companies cut people and waited for demand to recover. Today, many are cutting people while simultaneously investing in AI. That’s a different story, because AI isn’t just external noise anymore — it’s becoming a new operating logic.

For the worker, the difference between “it was the macro environment” and “it was AI” often doesn’t matter much. If your role is gone, it doesn’t make things easier to know which factor gets 60% of the blame. People don’t live inside analytical frameworks. They live in rented apartments, with children, mortgages, plans, and the question of what to do next.

AI as a convenient story for investors

AI has given managers a new vocabulary for layoffs. “We overestimated demand and need to cut costs” and “We’re becoming an AI-first company” can describe the exact same event. For the employee, the outcome is identical. For investors, it isn’t: the first version sounds like a mistake, the second sounds like a strategy. This is how AI-washing emerges — when artificial intelligence becomes cover for ordinary cost-cutting. The problem is that from the outside, these two situations often look almost the same. A company cuts staff. The company talks about AI. The company talks about efficiency. It’s easy for a journalist to write “company lays off workers due to AI” even when AI is just one of several factors, or simply a polished backdrop for an older story: growth is slowing and costs need to come down.

Someone is winning from this

Nearly every conversation about AI and jobs focuses on losses. But there’s another side. A small company that three years ago couldn’t afford a strong developer can today get by with one capable person using AI tools instead of two or three. The same dynamic is playing out in marketing, design, analytics, and finance. What used to require a small team can now be handled by one person with the right toolkit. For the labor market, that’s bad news. For small businesses, it’s often good news. The worker sees risk. The entrepreneur sees leverage. The investor sees margin. The customer sees a cheaper service. All of these perspectives can be true at the same time. But the most painful one socially is the first — because a worker can’t diversify their life the way an investor diversifies a portfolio.

What it means to be valuable now

The worst advice you can give someone today is “just learn AI.” It sounds right but means almost nothing. A more useful principle: value is shifting from executing simple tasks to being able to manage complex processes. Not just writing a text, but understanding what kind of text is needed, for whom, with what argument, and how to verify it doesn’t read like hollow machine output. Not just writing code, but understanding the system, its constraints, its architecture, its edge cases, and what will happen six months after the merge. Not just producing a report, but knowing which metrics are misleading and where the model confidently made something up.

AI performs well where the task is clear, repetitive, and sufficiently formalized. Humans become more valuable where the task is ambiguous, the stakes are high, context matters, and the cost of a mistake is significant. This isn’t a call to flee digital professions — but a career built solely on executing basic tasks is no longer a safe bet. That used to be enough to get started. Now it may not be enough to get in the door at all.