I read “The AI Layoff Trap” by Brett Hemenway Falk and Gerry Tsoukalas this week, and it’s stuck with me — not as an engineer, but as someone who’s spent the last few years building and leading teams through exactly this kind of transition.
The problem
Standard economic intuition says automation is a net good: it frees up labor, cuts costs, and the savings eventually flow back into the economy. The paper pokes a hole in that story. It asks a sharper question — why do individual firms lay off workers when adopting AI, even when doing so makes the overall economy worse off?
The core argument
The answer is a coordination problem, not a bug in any single firm’s logic:
- Each firm, acting alone, has a rational incentive to cut headcount once AI makes a role redundant.
- But workers are also consumers. Mass layoffs across many firms shrink aggregate demand.
- Shrunken demand hurts the very firms that just got “more efficient” — a self-reinforcing cycle the authors call the layoff trap.
It’s the classic gap between what’s individually rational and what’s collectively optimal — the same shape as a prisoner’s dilemma, just dressed up in AI-adoption clothing.
Why this matters to me
Leading a team that’s scaled from 4 to 32 people, my instinct has always been: AI should change what my team works on, not just how many of us there are. This paper gives some theoretical backing to that instinct — it suggests that treating AI adoption purely as a headcount-reduction lever is short-sighted even in narrow economic terms, not just a people-friendly platitude.
Takeaway
The paper argues this trap justifies policy intervention to smooth AI transitions. I’d add a practical, org-level version of that conclusion: teams that reinvest AI-driven productivity gains into new capability — rather than pure cost-cutting — are betting on the side of the argument that actually compounds.