A new research paper by economist Joshua Mask finds that new workers in remote work-exposed occupations are earning less than workers in other occupations. The same study finds no consistent negative effects in occupations with AI exposure. 

Young people entering the job market in 2026 are having a tough time. ZipRecruiter reports that entry-level jobs make up just 38.6% of job postings in March 2026—down from 43.4% in March 2023—and only one in four graduates are on their dream career path.   

Much of the blame for the weak job market has been focused on the arrival of AI and growing fears that the technology will replace human workers. But a new working paper from Joshua Mask, assistant professor of instruction in the Economics Department at the College of Liberal Arts, suggests we might be pointing our finger at the wrong culprit and that remote work may have a greater impact on new job seekers. 

“What I’m seeing across the board is that if you are in a remote work-exposed occupation and you entered the job market any time after 2020, you’re getting 2% lower pay,” Mask said. 

In his latest working paper, “Same Occupations, Different Clocks: Separating Remote Work from Generative AI for Labor-Market Entrants,” posted to Social Science Research Network on July 10, Mask finds that new workers in occupations with exposure to remote work are earning less than workers in other occupations.   

For the study, Mask sought to understand how remote work and AI were affecting the job market for new entrants. What made this tricky, he explained, is that jobs with exposure to remote work are largely the same as those with exposure to AI.  

Fortunately, there was a three-year gap between the spike in remote work caused by the pandemic and the public release of AI chatbots like ChatGPT. 

“Now that remote work and AI are firmly in place, it is hard to separate their effects,” Mask said. “But there was a period of time where there was no AI, but there was work from home. That three-year gap is largely what I rely on in this paper.” 

Mask used data collected by the Bureau of Labor Statistics through its Current Population Survey between 2012 and mid-2026 to analyze how AI and remote work affected new workers’ compensation, hours and employment. 

He found that new workers in occupations where work from home is common experience about 1.8% lower hourly pay and about 2% lower annual pay than workers in other occupations. 

He found no consistent negative effects in occupations with AI exposure. 

“The entry-level pay problem preceded AI,” Mask said. “It started with remote work. A lot of what people have been attributing to AI, like a fall in pay, may actually be because of work from home.” 

Mask can’t say from his research that remote work alone is responsible for the drop in pay. One possible explanation is what Mask calls a compensating differential. Essentially, workers may be choosing to take lower pay for a remote job because they save money working from home. 

“If you consider all the extra costs you incur simply by going to work, like buying gas and lunch, this gap may be that,” he said. “I’m finding a 2% lower annual income. On a $50,000 income, that’s $1,000. That may be your commuting cost.” 

Another explanation could be learning loss resulting from the pandemic. Remote work really took off in 2020, the same time that schools were closing down or going fully virtual. For all intents and purposes, many students lost a year of high school or college, Mask said. This may have led to a productivity loss when those students entered the job market.  

“If I didn’t learn Excel for my accounting degree, then I’m not going to be very good at my accounting job when I start off, and they’re probably going to pay me less, because my productivity is lower,” Mask said. 

Yet another explanation could be the end of zero interest rate policies and subsequent rate hikes in 2022. Mask explained that this could be a factor because many companies, particularly startups, rely on financing for their day-to-day operations. Rising interest rates may be affecting the same jobs that are remote and AI exposed, he said. 

“I can’t tell from this study whether remote work is at fault or whether it’s a combination of remote work and these other factors,” Mask said. 

As for why he’s not seeing the same drop in pay in occupations exposed to AI, Mask said that some concerns about AI affecting the job market may have been overblown. Mask wasn’t surprised by the finding, and he added that many economists wouldn’t be either. 

“I don’t know too many economists who have argued that AI is going to destroy jobs, mainly because that’s not what we’ve seen in the past,” he said. “Computers didn’t destroy jobs. The steam engine didn’t destroy jobs. Those things created more jobs.” 

Something the general public should consider, Mask says, is whether the negative effects they’re seeing in the job market are really caused by AI or whether remote work is to blame. 

“Everybody is worried about AI being a threat because they’re only seeing the beginning of it. It’s still very new, and people are looking at it through a tough lens,” he said. “What I’ve been telling my students more than anything else is to recognize that AI is probably going to be more of an opportunity than a threat.”