What is the 'dead economic theory' that claims AI-driven job cuts will devastate markets and shake democratic systems?
With the advancement of AI technology enabling AI to perform a variety of tasks, companies are scrambling to replace human employees with AI systems. Owen McGrand, who writes a blog on philosophy and technology, has dubbed the socioeconomic problems that AI-driven job cuts will bring about a 'dead economic theory,' and has pointed out the potential issues.
[2603.20617] The AI Layoff Trap
The Dead Economy Theory - by Owen McGrann - The Palimpsest
https://www.owenmcgrann.com/p/the-dead-economy-theory
Total investments in large-scale AI infrastructure by companies such as OpenAI, Anthropic, Google, Meta, and Microsoft have reached hundreds of billions of dollars (tens of trillions of yen), and are projected to reach trillions of dollars (hundreds of trillions of yen) in the next decade. OpenAI is valued at over $800 billion (approximately 12.7 trillion yen), and Anthropic is valued at nearly $1 trillion (approximately 160 trillion yen) despite never having achieved profitability. McGrun points out that a market large enough to justify these investments is needed, and that market is today's labor market.
Many white-collar workers are exploring efficient ways to use AI and experimenting to make their work more efficient. However, AI development companies are selling to investors that their AI agents can do the work of 10 analysts, essentially focusing on fundamentally replacing workers with AI and reducing staffing levels.
McGrun states, 'The financial model underlying AI requires the elimination of human cost centers on a civilizational scale. If that doesn't happen, these companies will become the most overvalued assets in capitalist history. People who write checks don't have the habit of wasting trillions of dollars on better autocomplete features or an endless increase in long notes that nobody will read.'
In fact, AI companies are publishing various benchmarks and research results to prove that they can replace human workers with AI. The GDPVal benchmark, devised by OpenAI, measures how well AI can perform tasks in professional fields such as lawyers, film directors, mechanical engineers, property managers, software developers, nurses, and pharmacists.
OpenAI develops 'GDPval,' a benchmark for measuring AI's real-world capabilities, enabling it to measure performance in real-world professions such as lawyers and film directors - GIGAZINE
In an article published in the New York Times in April 2026, the head of OpenAI stated that AI models were now achieving an 80% or higher win rate against human experts on tasks that no AI model could compete with just a few months prior. If we take this statement at face value, McGrand points out that the following changes would occur in society.
1: Corporate workforce reductions
Companies replace the majority of their workforce with AI, drastically reducing costs, increasing profit margins, boosting stock prices, and satisfying everyone attending earnings calls. In fact, when Block, a financial services provider , cut 40% of its workforce in March 2026 based on the assumption of using AI, the market reacted by seeing its stock price surge by 25.6% overnight. In other words, the market rewarded the elimination of human workers with a massive rise in stock prices.
2: Slump in consumer spending due to worker layoffs
Workers who are replaced by AI and laid off lose income and cut back on spending, which naturally leads to a decrease in sales for the companies that previously employed them. If only a small fraction of the vast number of companies implement AI-driven layoffs, the impact may be small, but as GDPval shows, the range of jobs that AI agents can replace is wide, so the impact could spread to the entire economy.
3: Market collapse
Companies that lay off workers in pursuit of their own profits will realize that many of their customers were also 'employees of other companies.' The more AI-driven job cuts spread across the market, the more consumer spending will decline and profit growth will stagnate. As a result, corporate investment in AI will lead to market collapse.
Professor Brett Hemmenway Falk and his colleagues at the Wharton School, the business school of the University of Pennsylvania, discussed this dynamic in their paper, 'The AI Layoff Trap.' In a competitive market, companies can enjoy all the profit increases from layoffs while bearing only a small portion of the resulting decrease in demand. Therefore, the incentive to 'lay off employees and cut costs' always outweighs the negative impact of 'decreased consumer spending,' leading every company to go beyond optimization and automate, ultimately causing the entire market to collapse.
Optimists in the technology industry argue that even if AI causes mass unemployment, new jobs adapted to the AI era will emerge, so there's no problem. In fact, in the United States, 90% of the workforce was once engaged in agriculture, but that percentage has now fallen to 2%. Furthermore, Professor David Autor of MIT
Certainly, these facts suggest that even if jobs are lost due to efficiency improvements and automation, new jobs will eventually emerge. However, the significant decline in the agricultural population occurred over a period of several generations, spanning 140 years, and does not align with the modern situation where a vast number of workers are losing their jobs due to AI in just a few years. Furthermore, Professor Carl Benedict Frey of Oxford University notes that it took 70 years for the wages and employment of workers who lost their jobs during the Industrial Revolution to recover.
While workers' wages stagnated during the Industrial Revolution, corporate profits increased, inequality worsened, and political instability arose. Frye stated, 'Most economists would acknowledge that technological progress can cause some adjustment problems in the short term. But what is often overlooked is that this 'short term' can last a lifetime.'
Daron Acemoglu , who won the Nobel Prize in Economics in 2024, concluded that the job losses caused by new technologies between 1987 and 2017 far outweighed the effects of increased productivity and re-employment, demonstrating that new jobs do not emerge as quickly as workers lose their jobs. Furthermore, while previous automations have replaced only parts of various jobs, such as 'looms' or 'spreadsheet software,' AI has the potential to simultaneously replace cognitive labor in all industries.
McGrun argues that AI replacement of workers is not only an economic crisis but also a political crisis for democratic systems. Democracy relies not only on the rulers setting up various systems and distributing resources, but also on the non-rulers providing labor, tax revenue, and consumer spending. He argues that the entire democratic system functions because there is value that can be provided to ordinary people who do not hold government or public power, and because power is distributed to a reasonable extent.
However, when the 'labor force' that ordinary people can provide is removed, and AI systems owned by a small number of large corporations accustomed to minimizing taxes begin to generate value, all financial mechanisms in a democratic system will become dysfunctional. The tax base will shrink, collective bargaining between companies and workers will become meaningless, and consumer spending funded by labor income will decrease drastically.
McGrun calls the problems caused by the reduction of workers due to the introduction of AI a 'dead economic theory.' McGrun said, 'A dead economy is not an economy where nothing happens. A lot will happen, and GDP may even rise. AI investments are already supporting that. A dead economy is an economy where a lot happens, but none of it needs you. The productive capacity of civilization is being taken away by a system in which you have no stake, no opinion, and no voting rights.'
Related Posts:
in AI, Posted by log1h_ik