The Meaning Externality: How AI Erodes Job Satisfaction Without Cutting Jobs
Economist Joshua Gans argues that the mere existence of a credible AI alternative can strip workers of their sense of purpose.
The threat of automation is typically measured in lost paychecks and unemployment rates, but a new theoretical framework suggests the damage begins long before a worker is fired. The psychological value of employment may be eroding simply because machines can now do the job.
In a working paper published by the National Bureau of Economic Research (NBER), economist Joshua Gans explores how automation harms workers even when they remain employed. Gans introduces the concept of a "meaning externality," arguing that the value workers derive from their labor is tied to the belief that their specific contribution is essential to the output. When a credible machine alternative exists, that sense of necessity vanishes, reducing the psychological reward of the work regardless of whether the human is still performing the task.
The Mechanics of Meaning
According to the research, this erosion of meaning is often triggered by the public demonstration of AI capabilities. A "meaning externality" occurs when an external developer proves a machine can perform a task, lowering the perceived value of the human alternative before a company even licenses the software.
Gans distinguishes between technical quality and public salience. While improving the technical quality of an AI tool increases the actual output of a firm, the public salience—the widespread awareness that the AI can do the job—is what weakens the meaning of human work. Essentially, the knowledge that one is replaceable is as damaging to the worker's psyche as the replacement itself.
Economic and Social Consequences
This shift moves the academic conversation beyond the binary of employment versus unemployment. The research suggests that the social cost of AI includes a systemic erosion of the "meaning of work," creating a workforce that feels alienated or undervalued despite having stable roles.
There is also a financial dimension to this psychological loss. The paper notes that if wages adjust fully to compensate for this loss of meaning, the firm bears the cost. However, if wages adjust only partially, the workers themselves bear the psychological and economic loss, effectively subsidizing the AI developer's innovation with their own well-being.
The Path Forward
This work builds on themes previously explored in "Power and Prediction: The Disruptive Economics of Artificial Intelligence" (2022), co-authored by Gans, Ajay Agrawal, and Avi Goldfarb. By shifting the focus to the qualitative experience of employment, the research provides a framework for understanding why AI development can be socially harmful even when it does not lead to immediate mass layoffs.
What remains to be seen is how firms will respond to this alienation. As AI capabilities become more salient, the industry may face a choice: increase compensation to offset the loss of meaning or risk a systemic decline in worker engagement and mental health across professional sectors.