Researchers studied AI layoffs. Here's their warning.

Overview: The “AI layoff trap”

Researchers and economists warn about an “AI layoff trap,” arguing that widespread AI adoption could trigger mass unemployment and an economy-wide demand collapse. Even if each company’s decision seems rational on its own, the aggregate outcome may be collectively harmful.

Core argument: the trap emerges under competition

The discussion centers on a paper—“AI layoff trap”—by economists Jerry Tsoukalas and Brett Fauske.

Monopoly vs. competition

• Simplified monopoly case: If one firm automates jobs, it recognizes that laid-off workers are also customers. This feedback makes the firm less likely to automate too aggressively.

• Competitive market case: Incentives shift:

  • Firms gain by cutting production costs with AI and can justify layoffs more easily because customers’ spending is distributed across many competing firms.

  • Each CEO reasons:

    “I should automate quickly or I’ll go bankrupt while others replace workers.”

  • If all firms act on this logic, aggregate demand erodes as a large share of the workforce loses income.

Why CEOs may still rush toward automation

The interview challenges the assumption that CEOs will simply “see the cliff coming” and avoid it. Even with awareness of macroeconomic risks, each firm’s immediate incentives can make rapid automation the dominant strategy:

• The cost of laying off workers may be low for an individual firm.
• The risk of being outcompeted by firms that automate is high.

This behavior is likened to game-theory dynamics such as the prisoner’s dilemma, where rational individual choices yield a collectively bad outcome.

Proposed solution: force restraint with policy

The paper argues companies will not voluntarily self-restain. It compares the policy challenge to Odysseus and the sirens, emphasizing that humans often need external constraints to avoid destructive traps.

Policy discussion and recommendation

The interview considers ideas such as:

• Universal basic income
• Worker equity

But it argues the most effective lever is one that changes incentives specifically around replacing workers. The main recommendation is an instrument akin to a:

• Tax on firing / full replacement with AI, so that:
  • augmenting workers with AI is relatively cheaper
  • fully replacing workers is more costly
  • firms are less likely to converge on coordinated mass layoffs

Historical and political practicality

The closest analogy to an enforced “externality mechanism” is described as a carbon tax (or related government targets). However, the speaker doubts that a straightforward tax would be politically easy.

Alternatives mentioned include:

• Subsidies for firms that retain workers rather than replace them.

Potential US future if the trap isn’t avoided

The interview suggests that if the trap is not mitigated, it could lead to:

• Extreme wealth inequality (described as unprecedented)
• Increased political instability, with a comparison to systemic shocks such as 2008

While the speaker stresses uncertainty in forecasting, they are confident that technological progress will continue, potentially intensifying the inequality risk.

Presenters or contributors

• Jerry Tsoukalas (economist; co-author of “AI layoff trap”)
• Brett Fauske (economist; co-author of “AI layoff trap”)
• Cady (interviewer)