"The first to state his case seems right, until another comes and cross-examines him." — Proverbs 18:17
There is a sentence you have heard so many times it has worn a groove in your brain: AI won't replace you — it will free you to focus on higher-value work. It appears in keynotes, in onboarding decks, in the press releases of the same companies shipping the tools. It is soothing by design. It tells you that the machine is a colleague, not a competitor, and that the future it is bringing is one of liberation rather than loss.
It is a testable claim. So let us test it — against the public record of what the companies making it actually did.
The Numbers Behind the Reassurance
In 2025, companies directly attributed roughly 55,000 job cuts to their adoption of AI — more than twelve times the number attributed to AI just two years earlier. Counting more broadly, over 100,000 workers were affected by AI-linked layoffs in 2025, and by partway through 2026 the figure for that year alone was already approaching 80,000. At least eight companies announced AI-related cuts of more than 10,000 people each.
The names are not obscure. Amazon eliminated roughly 14,000 corporate roles, explaining that AI lets it operate with "leaner structures" and innovate faster. Microsoft cut on the order of 15,000 jobs while positioning AI at the center of its productivity model. In May 2026, Meta cut around 8,000 roles — about a tenth of its workforce. By one tally, tech layoffs reached 142,000 in 2026 even as the same industry committed to something like $700 billion in AI infrastructure spending.
Set the slogan next to the data sheet. "AI will free you for higher-value work" was being said by leaders at companies that were, in the same quarter, freeing tens of thousands of people from their jobs. The liberation was real. It just ran in the other direction.
Where the Money Went Instead
Here is the detail that turns an awkward correlation into something sharper. When companies cut staff in the name of efficiency, you might expect the savings to show up as belt-tightening across the board — leaner everything, shared sacrifice. That is not what happened. As one financial analysis noted in 2026, companies were cutting jobs for AI while leaving untouched the billions flowing to investors through dividends and buybacks.
This is the tell. If the story were "we genuinely cannot afford these roles," the capital returned to shareholders would be the first thing to give. Instead, headcount gave, and the payouts held. The "efficiency" was not a survival measure. It was a transfer — value moved off the payroll and toward the balance sheet, with AI supplied as the narrative that made the transfer sound like innovation instead of a choice about who gets paid.
Even the premise deserves scrutiny. Not every economist accepts that AI is doing the displacing it's credited with. Fabian Stephany of the Oxford Internet Institute has pointed out that many firms over-hired during the pandemic and that some current cuts may simply be a delayed correction — with AI offered as the more flattering, more future-facing explanation. "We restructured because we hired too aggressively three years ago" does not move a stock. "We are an AI-first company building a leaner future" does. The technology is, among other things, a better story to tell about a layoff.
The Grammar of "Higher-Value Work"
Return to the phrase itself, because the phrasing is doing quiet work. "Frees you for higher-value work" contains an unstated second clause: and we will need fewer of you to do it.
If a tool makes each worker dramatically more productive, an organization has two options. It can keep its people and produce much more — growing into the new capacity. Or it can produce the same amount with far fewer people and pocket the difference. The slogan promises the first. The layoff announcements describe the second. Which one a given company chooses is not determined by the technology. It is determined by what the company is optimizing for, and the record is now fairly clear about what that usually is.
"Higher-value work" also quietly relocates the burden onto you. It implies that if you are displaced, the problem is that you failed to climb to the higher-value tier in time — a personal shortfall, not a structural decision made well above your pay grade. It converts a choice the company made into a deficiency you have. That is a remarkable piece of rhetorical engineering, and it is worth refusing.
What This Means If You Work for a Living
None of this means the tools are useless or that you should refuse to learn them. The opposite. It means you should hold a clear-eyed view of the deal you are actually in, and act accordingly.
Be the person who deploys the leverage, not the line item it replaces. The worker most exposed is the one who does a single, well-defined, automatable task. The worker most protected is the one who orchestrates the tools, owns the outcome, and carries the relationships and judgment that don't compress into a prompt. Move toward the second as fast as you can.
Own outcomes, not activities. Automation eats activities — the discrete, repeatable motions. It struggles with ownership: being the person accountable for a result, who decides what the result should even be. Tie your value to outcomes and you are far harder to "free."
Watch what your employer does with the savings. When a company adopts a productivity tool, the revealing question is what it does with the capacity it gains. Reinvest in growth, or return to shareholders and shrink? You are allowed to read that answer as information about your own future there.
Build something that's yours. The deepest protection against being treated as a fungible cost is to have value that doesn't live entirely inside someone else's org chart — your own audience, your own clients, your own body of work. The tools that make a giant "leaner" can make an independent operator genuinely powerful. The same leverage; a very different distribution of who it serves.
The Cross-Examination
The case for AI as liberation was stated first, confidently, and from every stage. It seemed right. Then the layoff notices arrived, and the dividends kept flowing, and the body count climbed into six figures while the infrastructure budgets climbed into twelve. That is the cross-examination, and the witness's story did not survive it intact.
The technology is not the villain here; it is genuinely capable of making work better. The thing to refuse is the narration — the soft sentence that asks you to experience someone's decision about your livelihood as a gift you were given. Learn the tools. Use the leverage. But do not let anyone tell you that being made redundant is the same thing as being set free.
Sources: Challenger and related 2025–2026 layoff tallies (55,000 AI-attributed cuts in 2025; 100,000+ affected; ~80,000 in 2026); CBS News on AI-cited layoffs at Amazon and others; reporting on Amazon's ~14,000 corporate cuts, Microsoft's ~15,000, and Meta's ~8,000 (May 2026); TechTimes on 142,000 tech layoffs in 2026 against ~$700B AI infrastructure spend; 24/7 Wall St. on job cuts alongside investor payouts; Fortune (March 2026) CFO survey; Fabian Stephany, Oxford Internet Institute, on AI as justification.
