Over the past year, “we’re automating it with AI” has become one of the most repeated phrases in corporate layoff communications. The narrative is seductive in its simplicity: artificial intelligence does the work, people are redundant. But reality, as almost always, is more uncomfortable and more interesting than the headline.
AI as Argument and as Excuse
It’s worth separating two things that often get mixed up. On one hand, there are tasks genuinely being automated: first-level support, routine report generation, certain documentation flows, and repetitive code. On the other hand, there’s the use of AI as a reputational justification for cuts that actually stem from prior over-hiring, margin pressure, or investor expectations.
Saying “we’re cutting because AI makes us more efficient” sounds like a forward-looking strategy. Saying “we over-hired during the boom and now we’re correcting” sounds like a management mistake. The same layoff, two stories. And the market rewards the first.
What the Numbers Hide
The underlying problem is that the productivity promised by AI still isn’t clearly reflected in the books. Several companies have discovered that the cost of running these systems at scale is far higher than expected: inference budgets burned through in months, licenses cut mid-year, teams tracking token usage as if it were fuel.
When a company lays off staff citing AI efficiency while simultaneously discovering that this AI costs a fortune in compute, the math stops being obvious. Savings on salaries can evaporate in GPU bills.
The Effect on Those Who Stay
There’s a less visible cost: that of the teams who survive the cuts. When a third of a workforce is eliminated under the promise that “AI will fill the gap,” the real burden falls on those who remain while the tools are still maturing. AI speeds up tasks, yes, but it rarely cleanly replaces an entire role. The result is usually more work per person, not less.
Why It’s a Powder Keg
The combination is explosive: layoffs justified with an argument the public perceives as inevitable, executives receiving bonuses while staff is cut, and a productivity promise that hasn’t yet materialized for most. Add to that the increasingly widespread feeling that anyone could be “automated” tomorrow.
That uncertainty erodes trust, complicates talent retention, and fuels a political debate about regulating automation that is only just beginning.
A More Sober Reading
AI will transform work, of that there is little doubt. But the real transformation will be more gradual and more nuanced than the “mass replacement” version dominating the headlines. The organizations that come out ahead won’t be the ones cutting fastest while invoking AI, but the ones that reorganize work so that people and systems complement each other.
In the meantime, it’s worth reading every “AI-driven” layoff announcement with a simple question: is this real automation, or a story that sounds better than the truth?
