Opening — The Brutal Numbers
489 people a day.
Not soldiers lost in war. Not patients in a pandemic. But the number of people in the tech industry who were laid off in 2025, on average, every single day. Over the full year, 180,000 jobs disappeared from the labor market.
And the numbers are not slowing down. In just the first few months of 2026, more than 150,000 positions had already been cut, with 20.4% — around 9,238 roles — explicitly labeled as "AI-related."
CEOs from Ford, Amazon, Salesforce, and JP Morgan Chase all seemed to say the same thing in unison: "white-collar jobs will disappear," as if they were reading from the same script.
But the question almost no one dares ask out loud is this: Can AI really replace human work yet? Or has it simply become the most elegant excuse for cost-cutting in modern corporate history?
HBR Exposes the Truth: People Are Being Fired for "Hope," Not "Results"
In January 2026, Harvard Business Review published research based on a survey of more than 1,000 executives. The results are the kind you have to read twice:
Only 2% of layoffs were linked to actual AI implementation.
Read that again: 2%
So what explains the other 98%? AI’s "potential" — not real, proven performance. These companies made anticipatory layoffs based on the belief that AI would replace people in the future, even though many had not deployed anything meaningful yet.
This is not a technological revolution. It is a narrative revolution. Companies are using the phrase "AI transformation" as cover for what they already wanted to do: cut labor costs without carrying the old stigma of traditional layoffs.
Think about it. If a company says, "We cut staff because profits are down," investors get nervous. But if it says, "We cut staff because we are transforming with AI," the stock goes up.
It is a win-win game for the C-suite — but a lose-lose for workers.
95% Invested and Got Nothing Back — So Who Pays the Price?
Data from MIT and Oxford reinforces what HBR found:
95% of companies investing in AI have not seen any return at all. That means roughly $30–40 billion in combined spending has vanished into pilot projects that never scaled, proof-of-concepts that never reached production, and AI tools employees never actually used.
So who pays for that failure?
Not the CEOs who approved the spending — they still collect bonuses for "leading AI transformation."
Not the boards that signed off on the budgets — they still get applause from investors excited by the AI story.
Employees pay the price — the people laid off because "AI is coming," even though AI still cannot truly do their jobs.
Data from the New York Federal Reserve confirms it: in the past six months, only 1% of companies in the service sector actually laid people off because of real AI deployment. The rest? Mostly restructuring with AI used as a convenient excuse.
Who Is Actually Losing Jobs? The Numbers Don’t Lie
If AI still cannot genuinely replace people, why are people still losing jobs? The answer is simple: people are losing jobs for real, but not because AI is already doing the work — because companies believe AI eventually will.
And the first people hit are not hard to identify.
By Job Function
- Administrative roles — face the highest risk, with 26% of tasks seen as replaceable by AI
- Customer service — follows at 20%
- White-collar work in general — 80% of American workers may have 10%+ of their tasks affected by LLMs (Large Language Models)
- Altogether, 25% of all current jobs may see AI take over portions of the work
By Gender — The Numbers Few People Talk About
This is the part mainstream coverage rarely highlights:
79% of working women in the U.S. are in roles at high risk from automation, compared with 58% of men.
Why? Because women are more heavily concentrated in jobs that AI is often seen as able to replace more easily — administrative work, data entry, customer service, and back-office functions — while men are more spread across physically intensive work or highly specialized technical roles.
The AI revolution is not gender-neutral. It is amplifying inequalities that already existed.
