Employees are navigating AI disruption in the dark

With the rise of generative AI (GenAI), organisations are moving quickly to restructure their workforces. Employees, though, often feel left in the dark by leadership—not fully understanding what those changes could mean for them. New research suggests GenAI could create a strategic risk for organisations as the gap between employer adaptation and employee anxiety steadily grows amidst reshaping workforces.

As the first major technology threat to white-collar work emerges, the evidence of employer adaptation is striking. My co-authors and I analysed job-posting data from more than 60,000 employers, tracking hiring patterns at S&P 1500 organisations across 13 years. This included data from before and after ChatGPT’s release in 2022. At organisations with the highest concentrations of white-collar employees, weekly job postings fell by as much as 18.9% after the release, a contraction comparable to hiring declines during the early months of the COVID-19 pandemic.

But the decline was selective. Organisations with stronger financial positions, greater digital readiness and more technically skilled workforces adapted most aggressively, cutting lower-level white-collar roles, such as administrative positions, junior analysts and entry-level office jobs, while accelerating demand for higher-level technical, managerial and AI-related talent.

Organisations and employees are dealing with GenAI’s disruption differently

Organisations are responding decisively and strategically to technological disruption. Yet within those same organisations, employees are reacting very differently. Glassdoor data from the same period shows significant declines in CEO approval ratings and business outlook scores. The sharpest drops came among junior employees whose roles were most directly affected by the hiring cuts. However, senior-level employees also reported feeling uneasy, suggesting the distrust runs deep.

When we examined the language employees used in their reviews, two concerns surfaced repeatedly: job insecurity and a lack of transparency from management. Without clear direction, many employees acted on their own. Among more than 15.8 million employees who changed jobs after ChatGPT’s release, those in vulnerable roles were more likely to pivot toward AI-related work or shift into less exposed occupations. Employees were not resisting change; they were navigating it alone.

READ MORE: Zuckerberg’s AI vision: More jobs, fewer workers and ‘personal superintelligence’

The issue is a fundamental corporate mismatch: Employers are executing a deliberate workforce strategy, while employees experience it as growing anxiety, uncertainty and declining trust in leadership.

Whether the gap reflects intentional withholding or poor communication, the consequences are the same. AI transformation requires employees to learn new tools, take on new responsibilities, and often redefine their roles. That kind of buy-in necessitates trust in leadership and confidence in where the organisation is headed.

Adaptation is uneven: Leaders’ response will determine if gap grows

The data also captured uneven adaptation. Women were significantly less likely to pivot to AI-related work, and R&D organisations, which might be expected to lead adaptation, cut hiring across levels even as demand for AI skills increased.

The research, detailed in a new brief from Columbia Business School’s Reuben Mark Initiative for Organisational Character and Leadership, offers a clear diagnosis. How leaders respond will determine whether this gap widens or closes. Transparent communication about which skills will be valued, how responsibilities will evolve and what the organisation is working toward may help close the gap revealed in the data. So, too, framing GenAI as a tool that expands what employees can do, rather than one that determines how many are needed, might help.

GenAI will reshape the white-collar workforce regardless of how any one organisation responds. What remains within leaders’ control is whether employees experience that transformation as an opportunity or as something being done to them. The difference will be determined not by technology, but by humans themselves.

About the Author: Wei Cai is an assistant professor in the Accounting Division at Columbia Business School. This article was co-authored by Philip G. Berger of the University of Chicago, Lin Qiu of Purdue University and Cindy Shen Xinyi of Northwestern University.

It was first published in HR Executive.