Published on 02-10-2025
When OpenAI introduced ChatGPT in 2022, the artificial intelligence (AI) platform quickly became a focal point in discussions about the future of work. Powered by a large language model (LLM), ChatGPT performs a wide range of tasks through prompt engineering, from coding and debugging to composing music, stories and essays. This versatility, combined with the rapid adoption of generative AI, positions the technology as a key driver in reshaping the labor market.
What distinguishes this iteration of generative AI from its predecessors is its proficiency in handling office jobs traditionally thought to be immune to automation. Characterized by their demand for high levels of education and skill, these roles were considered resistant to technological disruption. Nonetheless, generative AI exposes various white-collar occupations to a labor market shock, and despite its potential to massively influence employment, it remains unclear how employers and employees are responding.
In “Employer and Employee Responses to Generative AI: Early Evidence,” Daniels School Assistant Professor Lin Qiu and her coauthors Philip G. Berger at Chicago Booth School of Business, Wei Cai at Columbia Business School and Cindy Shen at Stanford Business School examine the reactions of both employers and employees to generative AI. The study directly sources daily job listings from over 60,000 employer websites and allows the tracking of real-time employer demand based on actual job openings rather than self-reported, aggregated data from different job boards. That allows for the identification of rapid shifts in employer demand triggered by the disruption of generative AI.
While past technologies primarily affected blue-collar jobs, the findings indicate that generative AI disproportionately affects white-collar jobs requiring high levels of critical thinking and creativity. The most exposed occupations include computer system engineers, writers and authors, and statistical assistants. These occupations generally involve high technical expertise, analytical abilities and effective writing skills, while the least exposed occupations mainly involve manual labor tasks. Moreover, the study finds that heightened exposure to generative AI does not inherently result in employee displacement, but can instead drive greater demand for high-level labor and management roles. Thus, heightened exposure to Generative AI may polarize the labor market by creating disparity between high- and low-requirement jobs.
The study also finds that firms with greater exposure to generative AI significantly increase emphasis on generative AI and machine learning skills in their job listings. Although employee ratings of firms’ current conditions remain stable, there is a notable decline in their long-term outlook about their firms. Additionally, potential employees seek fewer interviews at firms more exposed to generative AI, consistent with declines in hiring by such employers.
Given the positive responses of employers toward generative AI and the more negative responses of even those employees whose jobs are complemented by it, the results suggest mismatched responses toward the technology. The difference in perspectives could pose challenges in aligning organizational strategies with employee expectations, potentially affecting worker morale and trust in leadership.
Future research can build on this approach to explore real-time changes in internal firm dynamics in response to generative AI adoption. For example, these studies could examine how firms restructure roles and how they transition employees from lower-level tasks like bookkeeping to higher-level strategic or innovation roles. Other studies could look at how employer demand for AI-related skills shifts rapidly after new technologies are introduced, potentially exacerbating income inequality.
