Authors Brett Hemenway Falk and Gerry Tsoukalas’ recent paper, “The Layoff Trap,” examines how the economy would function if organizations displaced workers with AI faster than it could reabsorb them into the labor market. The premise is that replacing workers with AI would inevitably erode the very consumer demand firm’s depends on. It has often been postulated that new technologies displace workers within a company, but the same technology will create new jobs within the organization. For example, the advent of the internet changed the structure of many sectors, including the music industry, but also created demand for new skills. Physical recordings were replaced by digital distribution; the album format evolved into streaming; record labels’ vertical integration unbundled; and physical retail was replaced by digital distribution controlled by technology companies, including Apple, and by newly formed online music companies such as Spotify, all leading to new jobs, requiring new skills. In this scenario, the transitioning economy let each firm capture most of the cost savings from automation while bearing only a fraction of the demand loss it creates.
In their research, the authors argue that AI has turned technological labor practice into an automation arms race in which workers are displaced far beyond what is collectively optimal. The resulting loss harms both workers and firm owners. The more competition within a market, the more AI will amplify labor loss within that sector. Recent corporate restructurings in music and tech illustrate this shift, with firms cutting staff, partnering with AI platforms, and automating internal workflows. Studies from groups like the Music Workers’ Alliance report that most freelance session players and background composers have lost project-based income as brands and production houses use generative AI for initial demos and low-cost tracks.
Change is a natural part of organizational evolution in the entertainment industry. One theory, known as replacement theory or replacement investment theory, states that businesses should retire old assets, such as machinery, vehicles, or equipment, and replace them with new, more efficient models to remain competitive. However, this theory focuses solely on equipment replacement. It does not consider how these technological changes affect labor or the broader consequences of a reduced labor force on demand for goods in the economy. The scenario of organizations replacing labor with AI automation raises questions similar to those posed under “traditional” replacement models, including: who will buy goods if an organization automates every operation? For example, if one company makes a product and introduces AI, management must choose between laying off employees, who may be consumers, and keeping them to maintain demand. In a competitive market, a manager sees only a fraction of the market and will likely reduce staff, assuming the market will make up the difference. The problem is that if every company does this, demand erodes across the sector and all firms end up worse off or trapped.
The effect of multiple organizations acting in their best interest within a market, without regard to the overall quality of the environment, is known as a dominant strategy in economics. A dominant strategy occurs when an actor chooses an action that yields the highest payoff regardless of what the opponent does. When applied to human labor and employment, this game theory concept explains how employers and employees make strategic decisions about effort, wages, hiring, and retention. Without trust or incentives, both parties pursue their dominant strategies and end up in a Nash Equilibrium. The employer pays a low wage, and the worker provides low effort. Both sides end up worse off than if they had cooperated (high effort for a high wage), mimicking a classic Prisoner’s Dilemma in the workforce. The concept becomes especially useful for thinking about strategic behavior around generative AI in the music industry, because labels, artists, streaming platforms, and technology firms must make decisions while uncertain about what competitors will do. For example, two record labels—Label A and Label B—must decide whether to adopt AI or avoid it for activities such as production, marketing, recommendation systems, catalog analysis, or content generation. If Label B adopts AI, Label A should too because Label B may lower production costs, develop faster marketing tools, improve consumer analytics, create scalable content, or implement superior recommendation technologies. If Label B doesn’t adopt AI, Label A should, since it will give it advantages Label B doesn’t currently have. Even if they lay off the very employees who consume their products, music companies cannot afford to hold back on adopting AI within their organization.
As the marketplace reduces its workforce by replacing workers with AI, governments and economists are exploring several techniques to reduce labor loss’s impact on the economy. One example is for governments to provide a universal basic income (UBI). UBI supports demand and living standards, but it functions more like an automation tax than a general-revenue finance transfer, since displacement itself erodes general revenue. Nevertheless, raising living standards does not change any organization’s incentive to automate. Another approach the authors examine is a capital income tax, which taxes proceeds directly. As a tax on an organization’s profits, capital income taxes can be reinvested into the economy through public goods or debt service. This lowers the overall cost of capital and encourages companies to retain employees, freeing up cash flow to stabilize workforces during downturns. Governments can condition capital gains or corporate tax breaks on maintaining specific employment thresholds or payroll retention targets.
A market-based alternative to taxation is worker equity participation models that give employees a direct stake in the organization’s profits. This model aims not only to provide workers with income from their labor but, in theory, to incentivize them to invest a portion of their earnings in the organization itself, recycling capital back into the company. One issue with the solution is that delayed rewards can feel abstract. People manage bills on a monthly cycle; a yearly check acts like a gift rather than an active incentive. A similar model that incentivizes organizations to invest in their labor force directly is upskilling. This model is based on the traditional view that technology will provide new opportunities to an organization. Retrained employees not only provide the organization with skilled workers; they often move into higher-paying positions. Although AI optimists hold this view, evidence suggests AI-driven displacement would not create new labor-market opportunities, especially in an over-automated economy.
Examining models that do not require government intervention, the authors investigate applying Coasean bargaining theory to rectify labor loss due to AI. This economic model describes a process in which private parties negotiate to resolve conflicts over externalities (such as labor replacement by AI) and reach an efficient economic outcome without government intervention. Because automation is the dominant strategy, no voluntary agreement among firms is self-enforcing, and companies will prioritize efficiency over collaborating with competitors.
Finally, the authors explore the Pigouvian automation tax model. This theory centers on taxing market activity generated by organizations to counter the harm to society or individuals created by an organization’s activities, such as adopting technology that replaces labor. Just as carbon emissions impose hidden costs on the environment, displacing humans with AI imposes hidden social and fiscal costs, such as lost income tax revenue and increased demand for public assistance. As such, the tax endeavors to bridge the gap between corporate benefits and social costs. In some cases, the recommendation is to raise funds to support programs such as Universal basic income and to reduce public burdens. However, economist Ronald Coase identified three fundamental issues with this theory. First, Coates believes that externalities are fundamentally reciprocal. To cause harm, both parties must be present. Forcing only one side to pay can lead to an inefficient outcome where the cheaper “avoider” of the damage is not incentivized to act. Second, government intervention can reduce both parties’ incentive to negotiate an outcome and, in some cases, create market distortions and allocative inefficiencies. Finally, calculating the true cost of externalities is often difficult or impossible. Governments lack the precise market data required to set the tax properly. Because the regulatory state operates blindly, an incorrectly calculated Pigouvian tax can easily cause far more economic damage than the original externality itself.
As AI continues to be integrated into our everyday lives, music organizations, governments, and individuals must develop policies and procedures that ensure new technologies enhance our lives rather than diminish their quality. Although these issues not only pertain to the music industry, self-monitoring will not change the speed at which organizations will adopt AI. As investment in AI will continue to reshape market structure for the foreseeable future, the urgency for early government and sector policy action will become more prominent.
