The Impact of AI Adoption on Worker Productivity and Quality of Life
What if Michael Jordan Had to Mow His Own Lawn?
Imagine in 1996, if Michael Jordan of the NBA Bulls could not hire a gardener and had to devote 2 hours a day to mow his mansion’s lawn. His effort spent engaging in home improvement meant that he would lose time practicing with his team. Imagine if he could hire an AI robot to handle this task. Jordan would use this AI and reallocate his time to his core comparative advantage and he would become an even better worker and the Bulls would be more likely to win!
Let’s consider a less exciting example of how AI affects a worker’s time allocation. The typical Economics Professor has three main tasks; research, teaching and doing administrative work. Suppose that this professor works 11 hours a day and devotes 6 hours to research, 4 hours to teaching and grading and 1 hour to administrative work. Suppose that AI can handle the administrative work and 25% of the teaching grunt work such as emails and some grading tasks. This means that AI has saved this worker 2 hours a day that can be reallocated to research (or leisure!). This lucky AI enabled worker (the Professor) will enjoy a productivity boost because the least fun parts of the bundle of tasks that a Professor must deal with are handled by the Robot.
If AI also boosts the professor’s productivity (I.e Google searches yield more meaningful insights, the AI can format the paper and double check proofs and calculations), then each hour of research time generates greater quantity and quality of academic output. In this case, AI simultaneously reduces the “cost” of being an academic and raises the intellectual benefits of research time.
The examples of NBA Michael Jordan and the typical academic offer a preamble to discuss this new NBER paper.
Expertise
Working Paper 33941
Issue Date June 2025
When job tasks are automated, does this augment or diminish the value of labor in the tasks that remain? We argue the answer depends on whether removing tasks raises or reduces the expertise required for remaining non-automated tasks. Since the same task may be relatively expert in one occupation and inexpert in another, automation can simultaneously replace experts in some occupations while augmenting expertise in others. We propose a conceptual model of occupational task bundling that predicts that changing occupational expertise requirements have countervailing wage and employment effects: automation that decreases expertise requirements reduces wages but permits the entry of less expert workers; automation that raises requirements raises wages but reduces the set of qualified workers. We develop a novel, content-agnostic method for measuring job task expertise, and we use it to quantify changes in occupational expertise demands over four decades attributable to job task removal and addition. We document that automation has raised wages and reduced employment in occupations where it eliminated inexpert tasks, but lowered wages and increased employment in occupations where it eliminated expert tasks. These effects are distinct from—and in the case of employment, opposite to—the effects of changing task quantities. The expertise framework resolves the puzzle of why routine task automation has lowered employment but often raised wages in routine task-intensive occupations. It provides a general tool for analyzing how task automation and new task creation reshape the scarcity value of human expertise within and across occupations.
Here is a video of David Autor discussing his paper’s core ideas.
My Perspective
This paper explores the concept of "unbundling" at a broader level. Jobs consist of a variety of tasks, and very few employees are fully specialized in just one area. For example, consider a barista at Starbucks; she must be skilled at making hundreds of different drinks. Similarly, a postal worker needs to both drive a truck and walk to deliver the correct mail to the appropriate boxes.
Artificial intelligence can take over specific tasks where it has a distinct advantage. This allows workers to focus on the remaining tasks. If these tasks are enjoyable for the worker and are adequately compensated in the market, then the worker's job may improve as a result of AI.
Dynamics
At any given time, there is a fixed number of trained workers in the labor market. However, over time, some of these workers retire or leave their jobs, and new workers enter the field. As the impact of AI on various tasks becomes clearer, younger individuals are likely to focus on training for skills where AI can handle the more mundane aspects of work.
A concern arises if our nation’s public K-12 education system has not adequately prepared these young people to be adaptable and to seek out their advantages in the AI-driven modern labor market.
At USC Economics, I often ask my students, “What can you do better than a robot?” I want them to think critically about how they can position themselves to compete effectively in today’s economy.
A recent NBER paper highlights the crucial role of comparative advantage (and the Roy Model) in the labor market, illustrating how AI changes the economic benefits associated with specific skills. I am particularly interested in examining the short-term and medium-term effects of these shifts in the labor market.
In this brief Substack, I have not discussed whether AI will lead to overall wage compression or an increase in inequality. However, my discussions with professors indicate that, at least in the realm of academic output, AI is likely to exacerbate inequality. This is because high-performing individuals will have more time to think and create, which may result in even greater contributions to their fields.