New research from Washington University in St. Louis examines how workers’ perceptions of artificial intelligence (AI) have shaped their response to the emerging technology. The research also highlights how these differing responses could affect long-term career trajectories.
In a December 2025 survey of nearly 2,000 U.S. citizens, WashU graduate students Kiara Kim in Arts & Sciences, and Nathan Mester and Gregory Sun, in Olin Business School, found that whether employees adopt AI into their work largely depends on two factors: how they believe it will affect their current productivity and if they view it as a career-relevant learning tool — the latter being a more significant indicator.
The survey results are reported in a working paper, “AI and the labor market: A worker’s-eye view.“
“While much of the existing research has focused on how AI affects immediate output in specific jobs — such as call centers or software development — the workers in this survey appear to be thinking in longer horizons,” explained Yongseok Shin, the Douglass C. North Distinguished Professor in Economics in WashU Arts & Sciences, who supervised the students’ research.
“People are not only using AI to do tasks faster, but also to help them learn new skills, plan career moves or transition into new fields,” he said.
The majority of survey respondents indicated that they dedicate more time to learning now than before ChatGPT. This trend is driven primarily by those who view AI as a useful skill-building tool. Notably, this appears to be consistent across all age groups in the study, not just with younger employees — bucking the trend of employees spending less time learning as they age.
Additionally, the authors found that workers who view AI as a complement to their work have increased their effort at work since the introduction of AI. That’s because AI is generally used to automate the relatively easier tasks in one’s job, freeing workers to focus on harder tasks. This isn’t the case for workers who view the technology as a substitute for their work, though.
“Despite initial concerns that AI would replace workers, our research shows many employees are working harder than ever because they’re spending more of their time on the complex tasks,” said Kim, a doctoral student in economics.
Together, these findings suggest that AI’s ultimate impact on the labor market is not a mechanical consequence of its technical capabilities, the authors said. Instead, it emerges from the interaction between those capabilities and workers’ decisions about adoption, effort and learning.
How AI could reshape early careers, long-term success
‘Our research shows that forward-looking workers who increase both their effort and their learning in response to AI will get the biggest gains from the technology. ‘
Kiara Kim
The findings highlight the disadvantages that younger people face in today’s job market, according to Shin.
Many occupations follow an apprenticeship-like ladder: junior employees start with routine tasks, then graduate to more complex work as they gain experience. The survey confirmed that workers recognize this structure. Additionally, a majority agreed that routine tasks commonly performed by junior-level employees help build the skills needed for more advanced responsibilities.
By automating lower-level tasks that once served as on-the-job training, AI might disrupt the traditional career ladder, he said.
Exacerbating that disadvantage is the fact that experienced workers are aggressively training themselves using AI tools to prepare for career advancement, according to the survey.
Lifetime earnings are strongly tied to how much workers invest in their “human capital” by building skills over time. If AI disproportionately helps the already-motivated learners — those who see it as a complement to their own effort and a tool for long-term development — then AI could exacerbate earning gaps driven by these learning choices, the authors concluded.
“Our research shows that forward-looking workers who increase both their effort and their learning in response to AI will get the biggest gains from the technology,” Kim said. “Conversely, early-stage employees and those who fail to plug into AI-driven learning could find themselves not just behind in today’s tasks but structurally disadvantaged over their entire careers.”
The WashU Weidenbaum Center on the Economy, Government, and Public Policy provided grant funding for the research.