Replaceable but Employed: Automation and the Meaning of Work

# Replaceable but Employed: Automation and the Meaning of Work Source: [https://www.nber.org/papers/w35559](https://www.nber.org/papers/w35559) - [Home](https://www.nber.org/) - [Research](https://www.nber.org/research) - [Working Papers](https://www.nber.org/papers) - Replaceable but Employed: Automation… Working Paper35559 DOI10\.3386/w35559 Issue DateJuly 2026 Can automation harm workers without replacing them? We study jobs in which workers value both producing useful output and knowing that the output depends on their own contribution\. A credible machine alternative can weaken that second source of meaning even when the firm retains the worker\. Our model shows that this loss raises compensation when wages adjust fully; when they adjust only partly, workers bear some of the loss themselves\. It can also make automation more likely\. An external developer may profit by publicly demonstrating a machine before licensing it, because the demonstration lowers the value of the human alternative\. This "meaning externality" can create demand for the machine and make profitable development socially harmful\. Better technical quality and greater public salience have different effects: quality improves output, while salience alone weakens human work\. Automation can, therefore, reduce the value of work before it eliminates jobs\. - Copy Citation Joshua S\. Gans, "Replaceable but Employed: Automation and the Meaning of Work," NBER Working Paper 35559 \(2026\), https://doi\.org/10\.3386/w35559\. Download Citation ## Related ### Topics ### Programs ## More from the NBER  - Feldstein Lecture - Presenter:[Mark Duggan](https://www.nber.org/people/mark_duggan) Mark Duggan, who is currently the Wayne and Jodi Cooperman Professor of Economics at Stanford University and will soon\.\.\.  - Methods Lectures - Presenters:[Melissa Dell](https://www.nber.org/people/melissa_dell)&[Ashesh Rambachan](https://www.nber.org/people/ashesh_rambachan) Research Associate Melissa Dell of Harvard University and Ashesh Rambachan of MIT delivered the 2026 Methods\.\.\.