Companies have attributed some 80,000 layoffs in tech alone to AI year-to-date. Sure, just a fraction of the 8.8 million jobs lost in the broader financial crisis that followed the collapse of Lehman Brothers in 2008. But still a significant number. And one that's growing.
No one has enough visibility into each of those businesses to judge whether AI was truly the primary driver. Or whether "AI" is being used to repackage older causes: overhiring, margin pressure, industry dynamics, or plain mismanagement.
But one thing is clear, I think: pinning the blame for every unpleasant outcome on AI teaches the public what AI means.
And the companies don’t even need to label this as replacing "lower-value human capital" with AI, as Standard Chartered's CEO so subtly did recently.
If AI is consistently introduced to people as the punchline to a redundancy email, it stops being "the next productivity leap" and starts being a shorthand for corporate extraction.
The numbers bear this out. Only 26% of voters report positive feelings about AI, compared with 46% who hold negative views. And you don't need to be an undergraduate to connect the dots: if you already feel entry-level paths are disappearing, "AI did it" becomes an intuitive response, whether it's accurate or not.
That's why "AI made us do it" works better for shareholders than for everyone else. Investors hear efficiency, inevitability, and competitive discipline. Employees, customers, and voters hear: you are replaceable, and nobody is accountable.
Two (both bad) scenarios may follow:
1. The cry-wolf scenario
We continue labeling today's restructurings as "AI-driven" while the technology is still, in reality, a clumsy project. It's useful and impressive, but uneven and brittle, and still heavily shaped by human choices.
Then, when genuinely high-stakes AI risks show up at scale, warnings will be ignored. People will shrug. Nobody will believe the messengers any longer.
2. The backlash scenario
If AI becomes the universal excuse for job loss, it also becomes a universal target. Anger that should be directed at governance and incentives will get redirected toward the technology itself.
Both treat AI as fate, when it is (still) a choice. The alternative is treating automation as an implementation decision that companies own, not an inevitability they're powerless to resist. And only crying wolf when the wolf is actually at the door.