AI layoffs are rapidly becoming one of the biggest business stories of the decade. Every week, companies announce workforce reductions tied to artificial intelligence, automation, and digital transformation. But an uncomfortable question is emerging: are these AI layoffs truly driven by technology—or is AI becoming a convenient excuse for decisions that were already going to happen?
A recent court ruling in Hangzhou, China may have exposed a much larger issue hiding beneath today’s AI transformation narrative. Look right away at the irony of this story: recently, A tech company in Hangzhou, China’s eastern AI hub, deployed AI large language models to take over a quality assurance supervisor’s core responsibilities The supervisor’s job — his core responsibility — was exactly verifying the accuracy of AI-generated outputs. That’s right. The man whose job was making sure AI didn’t mess things up… was fired because AI took over his job of making sure AI didn’t mess things up.
But here’s where it gets really interesting. The company didn’t fire him right away. They first offered him a “reassignment” — a quieter way of saying demotion — at a 40% salary reduction. When the employee, identified only by his surname Zhou, refused, the company terminated his contract. Their justification? Organizational restructuring and reduced staffing needs driven by AI adoption.
Zhou fought back. He filed an arbitration claim, and eventually He won. But as the company sued to overturn that decision, it eventually lost at the district court level, and And just last week, they lost again on appeal.
Why so? The Hangzhou Intermediate People’s Court ruled the dismissal unlawful on two grounds.
First, AI-driven workforce reduction does not constitute a “major change in objective circumstances” under China’s Labor Contract Law — the legal threshold required to justify termination based on redundancy.
Second, the steep salary cut embedded in the reassignment offer was itself unreasonable. The court’s conclusion was as elegant as it was radical: companies cannot shift the costs of technological transformation onto their employees.
And here’s the reflection that lies behind all of this: If this ruling becomes the global template, companies will have to prove that the AI actually does the job before firing the human, and that’s a fascinating forcing function.
Right now, “AI did it” is a story executives tell shareholders to make a margin trade sound like a transformation, but a legal regime that demands evidence would expose how many of these so-called transformations are accounting tricks dressed up as innovation.
And as someone who’s sat on the executive side of the table where these decisions get made let me tell you, the uncomfortable question for leaders is this: would your AI layoffs survive a courtroom?
Especially right now. Because we’re watching something strange unfold in real time. Companies are firing thousands of people “because of AI.” And then, quietly, they’re hiring people back. I’d call this the AI layoff boomerang. So which is it: are these layoffs for real — or are they just an alibi?
That’s what we’re going to talk about in this article.
The world is quietly redrawing the rules
Let’s zoom out for a second. See, the Hangzhou ruling is not an isolated development.
It builds on a December 2024 arbitration decision in Beijing involving a data mapping worker whose job was automated away. That panel reached the same conclusion: an employer’s decision to adopt AI is not an uncontrollable event, and the costs of that business choice cannot be unilaterally passed to employees through dismissal or demotion.
And it’s not just China. Watch what’s happening, at the same time, on three other continents.
In the European Union, mass layoffs driven by automation can trigger consultation obligations under the Collective Redundancies Directive. Translation: you have to sit down with works councils and explain yourself before you cut. And starting in late 2027 — after the recently agreed Omnibus deferral — the EU AI Act will add a separate compliance layer for high-risk AI systems used in employment, including hiring, performance management, and task allocation. With penalties up to 35 million euros or 7% of global revenue. That’s GDPR-level teeth.
In the United Kingdom, dismissals linked to technological change can engage unfair dismissal protections. Tribunals look at whether employers followed fair processes and genuinely considered alternatives.
And across Asia-Pacific — in places like Singapore and Japan — you have mandatory reemployment obligations, notice periods tied to seniority, and sector-specific labor rules. All of which create real legal exposure for employers who think AI-related restructuring is just a clean business decision.
So what does all this add up to?
Three philosophies of what an employer actually owes a worker. Same technology, but three completely different answers. China says the employer bears the cost of transformation, period. Europe says you have to consult, document, justify. The United States… the United States says “AI did it” is a sufficient explanation, and at-will employment does the rest. Interesting as China is at the same time at the forefront of the AI revolution!
And here’s what every multinational leader is about to discover: your “values” — that beautiful slide at the top of every culture deck — will now produce different outcomes depending on which country your worker happens to be sitting in.
The Boomerang Effect
Now, opponents of regulation will surely call all of this public interference. The state getting in the way of innovation. The visible hand strangling Adam Smith’s invisible one. And would obviously argue that this is the wrong way to go, and that the US got it right.
But here’s the thing that’s interesting. The Chinese court isn’t ahead of the market. The market is finally catching up to the court.
What we’re calling a legal restriction is actually the market quietly conceding that full replacement was also an error, and eventually backfired in a number of cases.
Let’s take Klarna as an example — the Swedish fintech that became the poster child of AI replacement — fired roughly 700 customer service workers and replaced them with an OpenAI-powered chatbot. CEO Sebastian Siemiatkowski went on CNBC to proudly share that Headcount dropped from 5,500 to about 3,400. Investors applauded. Then six months later, customer satisfaction collapsed. The chatbot looped people in circles. It couldn’t handle emotionally charged interactions, disputes, the messy 20% of cases that actually matter. Klarna started quietly rehiring, and Siemiatkowski said something remarkable: “From a brand perspective, a company perspective, I just think it’s so critical that you are clear to your customer that there will always be a human if you want.”
Translation: we went too far. Maybe it’s not a coincidence now that Klarna reached break even for its first time on May 2026.
IBM did the same with HR. The company cut 8,000 HR jobs and pushed them into a system called AskHR. Then it discovered the AI couldn’t handle empathy, judgment, or any case that didn’t fit the script. So it started rehiring humans.
And then there’s Duolingo — and this one is interesting, because it’s a softer version of the same pattern. In April 2025, CEO Luis von Ahn sent his now-famous “AI-first” memo: contractors would be phased out where AI could handle the work, and employees would be evaluated on how much they used AI. The framing went viral — and not in the way he wanted.
A year later, von Ahn has quietly walked back almost every piece of it. The AI-usage performance metric? Gone — after his own employees asked whether they were just being told to “use AI for AI’s sake.” The “AI-first” rhetoric? Softened. In a recent Fast Company interview, he admitted: “I did not give enough context.” And he was even more candid about the technology itself: “The happy path is really fast. But the unhappy path makes it so that you end up spending more effort on that than the time you saved on the other thing.”
In plain language: AI didn’t deliver what the memo promised. So the memo got rewritten. One walk-back at a time.
See the pattern? Klarna fired and rehired. IBM cut and rehired. Duolingo declared and retreated.
And these are not isolated cases. Forrester’s Predictions 2026 report found that 55% of employers now regret laying off workers for AI, and Gartner predicts that half of all AI-attributed layoffs will be reversed by 2027 — though, and this part stings, often offshore or at significantly lower salaries.
So is this a boomerang effect? It is. And it’s a particular kind of boomerang, because the version of you that comes back is not the version that was fired. It comes back lower-paid, less stable, often in a different country, with less institutional knowledge, less morale, and more skepticism about whether their employer ever actually understood what they did for a living.
Now, let’s go even deeper, because the more I think about this boomerang effect, the more I ask myself: Because if the market is having to walk back its AI layoffs — what does that say about how real those AI layoffs were in the first place?
This brings us to what Sam Altman himself has now publicly acknowledged. He calls it “AI washing.” And the data is striking.
Forrester’s Predictions 2026 report found that nearly 6 in 10 hiring managers admit AI was cited as the reason for layoffs that were actually driven by budget cuts, revenue uncertainty, or the need to unwind the aggressive hiring sprees of 2021 and 2022. Six in ten.
Nikkei Asia attributed about 48% of Q1 2026 tech layoffs to AI and automation. RationalFX put the explicit AI-attribution figure at 20% for the same period. That’s a gap of nearly 30 percentage points between two analyses of the same quarter. Why? Because companies escalated the AI framing of their cuts as the quarter progressed — not because the underlying cause changed.
In other words: the AI didn’t take more jobs. The narrative just got more convenient.
Altman put it plainly at the India AI Impact Summit: “There’s some AI washing where people are blaming AI for layoffs that they would otherwise do.” When the CEO of OpenAI is the one telling you the AI replacement narrative is being abused, you should probably listen.
So why are companies doing this? Three reasons.
First, AI layoffs play differently on Wall Street. Marc Benioff said it best when he announced 4,000 customer support cuts at Salesforce: “I need less heads.” Layoffs framed as “cost cutting” trigger questions about why you hired too many people in the first place. Layoffs framed as “AI transformation” trigger a 3% stock pop and a Bloomberg piece about leadership.
Second, AI gives executives moral cover. If you say “we’re cutting because of AI,” you’re not the villain. AI is the villain. AI is weather. It’s just happening. Nobody made a decision, but in reality it’s your human decision outsourcing responsibility to AI. Convenient, isn’t it?
Third, it’s contagious. The clustering of Oracle, Meta, and Snap layoff announcements in the first three weeks of April 2026 wasn’t coincidence — it was a corporate playbook becoming standard practice. Once one CEO frames cuts as AI-driven and gets rewarded, every other CEO has to ask whether not doing so makes them look behind.
So what we have, in plain language, is a market where companies are firing people for a story their CFOs don’t fully believe, hiring them back when the story doesn’t work, and reframing the whole sequence as innovation.
And here’s the deeper issue. There’s a real cost to this kind of dishonesty, and the cost is trust. The Chinese court’s ruling, the Klarna reversal, the AI washing data — they’re all chapters in the same story, a story that a generation of workers is figuring out that “AI” was the name we gave to a decision that was made by a person. With a face. And a title.
Now pause here for a second. I want you to do something.
Think about the last time you read a headline that said “Company X is cutting Y thousand jobs as part of its AI transformation.”
What did you feel?
If you felt sympathy for the workers and a kind of fatalistic respect for the company’s “courage” to “embrace the future” — congratulations. You just did exactly what the press release was designed to make you do.
Now imagine the same headline read instead: “Company X’s CFO decided to cut Y thousand jobs to hit a quarterly margin target, and used AI as the explanation.”
Same facts. Different story. Completely different ethical weight.
The grammar of a layoff matters. Because it’s the difference between something that happened to people and something that was done to people.
AI is a choice, not a force of nature
This brings me to what I think is the deepest layer in this whole story. And honestly, it’s the reason I wrote this article at all.
The Hangzhou court made a legal argument. But buried inside that legal argument is a philosophical one — and I’d argue it’s one of the most important reframes of the AI conversation we’ve seen in years.
The court said AI-driven workforce reduction does NOT qualify as a “major change in objective circumstances.” Now, “objective circumstances” sounds like legal jargon. But unpack it. What the court was really saying is: AI is not weather. It’s not a flood. It’s not a market collapse. It’s not an act of God. It’s a choice a company makes. And because it’s a choice, the cost belongs to whoever made the choice. Not to the worker who was reorganized around it.
I want you to sit with that for a second. Because every layoff narrative of the last two years has leaned on the same trick. AI is treated as something happening to the company. “The technology is moving so fast.” “We have no choice but to adapt.” “The future is here.” Notice the passive voice. Notice the disappearance of responsibility?
The Chinese court called the bluff. And the deeper question isn’t legal — it’s grammatical. When we say “AI is replacing workers,” who is the subject of that sentence? Strip the passive voice away and the moral weight reattaches to whoever signed the org chart.
An AI layoff is nothing more than a managerial decision wearing technological clothing.
This is the heart of one of the nine skills I lay out in my book Between You and AI. I call it Agency. And I’d argue it’s the most important skill in the entire framework — the one that ties the other eight together. Agency means owning your decisions, even when AI is involved. Especially when AI is involved.
Because here’s what happens when leaders surrender their agency to “the AI”: they don’t just dodge accountability for the bad calls. They also lose the ability to make good ones. If you tell yourself the technology is forcing the layoffs, you stop asking the harder questions. Was this a transformation, or was it a cost cut? Did we even understand what work this person was doing? Did we test the AI on the messy 20% before we eliminated the human who handled it? Are we confusing tasks with jobs?
The leaders who get this right in the next five years won’t be the ones who automate fastest. They’ll be the ones who own their choices clearly enough to make better ones.
Because if you can’t say out loud — to your team, to your board, to the worker you’re letting go — “I decided to do this, and here’s why” — then maybe you shouldn’t be doing it.
So let’s come back to where we started.
Somewhere in Hangzhou, there’s a man named Zhou. We don’t know his first name. We only know he was a quality assurance supervisor whose job was to make sure AI got things right.
And his employer decided that AI no longer needed his supervision. They offered him 60% of his old salary or the door. He chose neither. He chose the courts. And the courts told his employer something that every executive listening to this should hear.
You don’t get to call your decision the future. You don’t get to call your decision the technology. You don’t get to call your decision anything other than what it is — your decision.
The question for the next decade of leadership isn’t going to be whether AI takes jobs. It’s going to be whether you, as a leader, can stand behind what you do with AI, without hiding behind it.
So I’ll leave you with this. When you make your next big AI call — the next restructure, the next automation rollout, the next “AI-driven efficiency play” — picture the courtroom. Picture Zhou. Picture the version of your decision that has to survive contact with someone whose life it just changed.
If your decision can’t survive that picture, maybe it isn’t really about AI.
Want to continue exploring the impact of artificial intelligence, leadership, and the future of work? Discover my podcast Between You and AI | Leadership, Human Skills & the Future of Work.
Check out my book “Between You and AI,” published by Wiley — that’s where I unpack the whole framework of nine human skills for navigating this wave. And you can find me at andreaiorio.com for everything else: keynotes, content, conversations.
