The conversation about artificial intelligence (AI) has often focused on what jobs it might replace—or how much time it could save in the workplace.
But new research by Wei Jiang, vice dean of faculty and research, and Charles Howard Candler Professor of Finance, suggests AI may be doing the opposite: extending the workday.
“AI and the Extended Workday: Productivity, Contracting Efficiency, and Distribution of Rents” is a National Bureau of Economic Research working paper. Jiang is a co-author alongside faculty from Auburn University, Fordham University, and Seton Hall University.
The group found that when someone transitions to a job with high exposure to AI, they work 3.5 hours longer per week. This imbalance ends up replacing non-screen-based activities, such as socialization and exercise.
“We are the only research so far to focus on the intensive margin,” explains Jiang.
“Everyone else is talking about what jobs AI will replace, but they’re not talking about how people work when they’re on the job.”
The Road to Research
Jiang is especially interested in disruptive AI. Think Microsoft Word installing a word-finishing software in 2002, which prompts users with suggested words once they start typing. In 2021, the technology behind real-time lane detection was patented. In 2023, Google began experimenting with its AI summary feature.
When ChatGPT was released, Jiang was excited about the time she would save with her new “assistant,” she jokes, but she quickly realized that it was taking more time to catch up with work. She began to wonder if others were experiencing the same phenomena.
She conducted an informal survey of people from various professions that may have been impacted by AI, such as medical doctors, lawyers, and investment bankers. She found that they were also working longer hours.
“As a researcher, whenever I feel something is going on, my next question is, ‘What kind of data can I find to prove it right or prove it wrong?’” says Jiang.
Behind the Research: The Method
Around the same time, Jiang was reading a book that referenced the American Time Use Survey, a database rich with details about how Americans spend their days, broken down into 15-minute increments.
Jiang and her co-authors began researching this data from the survey, as well as abstracts from the most disruptive AI patents and job descriptions of every occupation maintained by the U.S. Department of Labor.
Comparing the patents’ abstracts with job descriptions, the researchers were able to focus on jobs with high exposure to AI. Using the time use data, the team discovered an almost 10 percent increase in the length of the workday in these jobs.
The “principal/agent” framework is another component of the research, which demonstrated increased working hours when jobs have the potential to be monitored by AI. Think of the manager as the principal and a staff member as the agent; the CEO as the principal, and the manager as the agent—and so on.
The more AI becomes proficient in helping ‘principals’ monitor delegated work, the longer ‘agents’ work.
Wei Jiang, Vice Dean of Faculty and Research, and Charles Howard Candler Professor of Finance
Given this framework and the overall conversation about AI replacing people, it’s easier to see how people would work longer hours with AI. The perspective is often humans versus AI, says Jiang, when it should be humans plus AI.
In a different paper, Jiang compares humans plus AI to centaurs, the mythical creature that’s part human and part horse.
“We want to find a fulfilling place for humans—that’s the whole purpose of why we created AI, but humans will no longer be simply humans; they will be centaurs,” says Jiang.
The Future Life of Workers With AI
With the advent of new technology, it’s not uncommon for society to experience something called the leverage paradox, explains Jiang. This concept states that with a lever, a task should be made easier. With AI, life should be easier, for example.
The paradox is that it’s not the case: The task is made more challenging. With AI, work could last longer.
“I’m cautiously optimistic. It feels like we are in a transition of this leverage paradox,” says Jiang.
Eventually people have this reckoning and try to line up the original purpose of this technology to the adoption, but in transition, the instant move is people work more to stay competitive.
Wei Jiang
This competition stems from an imbalance: “Suppose I’m positioned at this point of the job market, but when new technology comes, there’s a reshuffling, and suddenly I don’t know where I belong in the spectrum. So, you have to pedal extra hard to get ahead or maintain your position.” It explains the longer working hours, particularly with regard to the potential for AI surveillance.
In fact, looking at additional databases, including Glassdoor and LinkedIn, the team found that when the labor market is more competitive, the length in work hours is more pronounced.
Right now, the power lies with the technology and the principals (the employers). Once society catches up—as people can only work so many hours in a day—the power will return to the people.
Jiang uses electricity as an example: When lights were invented, factory owners believed workers could stay in factories after dark, working longer hours, until laws caught up and the workday was shortened.
Today, people experience burnout and job dissatisfaction, and companies may experience increased turnover.
“When the AI exposure is higher, the job satisfaction is lower,” explains Jiang, citing a finding from the Glassdoor data.
The Researcher
Jiang began her academic career after earning her doctorate in economics from the University of Chicago in 2001. She found a home at Columbia University’s business school and worked there for 21 years before joining Goizueta’s team in 2022. Despite the longevity of her career, Jiang says she has never been more productive than since she joined Goizueta’s faculty.
“My three years at Emory have been the most productive years in my whole career in terms of papers published in top journals, research excellence awards I’ve received in the profession, and keynote speeches I’ve delivered around the world,” says Jiang, who is the incoming president of the American Finance Association.
As vice dean of faculty and research at Goizueta, a dean-appointed position, Jiang is exposed to not only the research happening within Goizueta’s walls, but also the broader Emory campus.
A university is a place where knowledge is created and disseminated, and faculty play a central role in both creating and disseminating knowledge. Being in a position to support and promote such a process is very satisfying.
Wei Jiang
Jiang has published two other AI-related research papers, both of which received “annual best paper” from their respective journals, since joining Goizueta. One paper, titled “How to Talk When a Machine Is Listening: Corporate Disclosure in the Age of AI,” covered corporate disclosures. Published in The Review of Financial Studies, the paper shows the impact of AI reading these materials for human investors, stock analysts, and lawyers. The other paper, titled “From Man vs. Machine to Man + Machine: The Art and AI of Stock Analyses,” discusses how human stock analysts could use AI to beat AI; it was published in the Journal of Financial Economics.
“My [research] theme is what humans do with technology to survive the disruption,” says Jiang. “I’m especially interested in how AI is shaking the foundation of how we govern the business world.”
Jiang approaches her research findings by inviting others to draw their own conclusions from the evidence she presents with the goal of opening doors and starting critical conversations. With the findings from “AI and the Extended Workday,” Jiang hopes companies will see that while there is a current productivity gain, the long-term will bring burnout, lack of motivation, and turnover—until a healthier work-life balance is attained.
From artificial intelligence to finance, Goizueta faculty are advancing research that redefines how we work and lead. Discover more.
