AI and Automation: An Economic Trap for Businesses

AI Automation: A Threat to Economic Demand

A study conducted in March 2026 by Brett Hemenway Falk and Gerry Tsoukalas highlights the potential dangers of automation through artificial intelligence (AI). According to their research, this automation could destroy jobs at a faster rate than they are created, which would erode demand and destabilize macroeconomic balance.

In a context of fragmented competition, each company benefits from cost reductions due to automation, but this comes at the expense of overall consumption. This phenomenon generates a negative externality, where widespread automation leads to a decrease in profits. The authors of the study suggest implementing a Pigovian tax on each automated task to internalize the costs associated with the decline in demand and achieve an optimal economic balance.

Productivity and Economic Trap

Automation is often seen as a means to increase productivity. However, a recent study warns of an economic trap where AI, by destroying jobs, ultimately reduces demand. While there is almost unanimous consensus on the impact of AI on the economy and employment, opinions differ on the manner and speed of this transformation. Some optimists believe that AI, like any technological revolution, will create new jobs, while others fear it will undermine the foundations of our economic model.

The Impact of Automation on Demand

Announcements of job cuts due to AI are multiplying, even in sectors like journalism. This accumulation of decisions could have significant macroeconomic effects. If AI replaces workers faster than new jobs are created, household incomes decline, leading to a decrease in overall consumption. This aggregate demand is essential for the sales of all companies. When labor income decreases, part of the customer base disappears. In theory, every company is aware of this interdependence, but in practice, it only bears a limited fraction of the contraction in demand while benefiting from cost-cutting gains.

The Prisoner's Dilemma and Automation

This competitive mechanism resembles the prisoner's dilemma in game theory. In an ideal scenario, all tasks are easy to automate, and companies can adjust their level of automation without constraint. Each company must choose between automating to reduce costs or slowing down to preserve jobs and thus demand. Individually, the dominant strategy is automation. However, collectively, this leads to a lose-lose situation. By simultaneously reducing employment, companies weaken their own consumer base, resulting in a decrease in overall profits.

Solutions to Automation

The central question remains: can this mechanism correct itself, or does it require public intervention? Researchers emphasize that it is legitimate to ask whether public policies can correct this dynamic. The study examines several often-proposed responses, such as increasing competition, wage adjustments, greater employee participation, basic income, training, and traditional profit taxation. However, none of these solutions are sufficient to halt the dynamics of over-automation. The authors advocate for a Pigovian tax on each automated task, aimed at internalizing negative externalities. By increasing the cost of automation, this tax would encourage companies to replace human labor only when productivity gains genuinely offset the loss of purchasing power. This mechanism would help align the level of automation with a social optimum, where neither companies nor workers would be sustainably penalized.