Workers who fear AI will take their jobs are likelier to justify slacking off, and the effect was stronger in finance and insurance

Washington State University psychologist Tahira Probst and her co-authors surveyed workers in 18 industries and found that people anxious about their occupation’s future were more likely to say they justified behavior they would normally call improper, and more likely to report slacking behaviors such as web surfing and showing up late. The pattern was a statistical association in survey responses, not a demonstrated cause of any individual’s conduct.

Make AI work for you: The Neuron delivers the latest AI tools, tutorials and news straight to your inbox. Join 700,000+ readers and catch up in 3 minutes a day.

Get The Neuron →

We may earn a commission if you sign up.

The paper, published September 9, 2026, in Occupational Health Science, carries a title that signals its scope: “In the Shadow of AI: How Industry Exposure to AI and Occupation Insecurity Fuel Moral Disengagement and Counterproductive Work Behaviors.” Probst is a professor of psychology at WSU Vancouver and the lead author. Her co-authors include graduate student Rebecca Lindgren, also at WSU Vancouver, and researchers based in Italy and Belgium.

The chain from job insecurity to moral disengagement

The study’s central model, set out in the paper, runs in three steps. Workers who feel insecure about their occupation because of AI report higher moral disengagement, and higher moral disengagement goes along with more self-reported counterproductive work behavior. In the WSU news release, distributed through EurekAlert, Probst defined the middle step in plain terms: “Moral disengagement is basically a way of justifying bad behavior.”

Counterproductive work behavior is the formal label for conduct that works against an employer or colleagues. The release gives web surfing during work time and arriving late as examples, and the authors suggest that some employees may rationalize cutting corners or undermining others as a way of protecting themselves or evening out a perceived injustice. Probst also noted that such behavior damages the working relationships inside an organization: “Counterproductive workplace behaviors are harmful to the interpersonal relationships within an organization,” she said.

All of this is measured by what workers say about themselves. The release does not report how many people answered the survey, the response rate or the size of the statistical effects, and the full paper at Springer is where those details would sit. Self-report also cuts two ways: it captures conduct that managers never see, but it depends on respondents describing their own lapses honestly.

How the researchers sorted industries by AI exposure

To go beyond individual worries, the team used the AI Industry Exposure index, an existing database that ranks industries by how much AI disruption they could face. That ranking is an objective measure built independently of the survey, so the researchers could ask whether the same worry produced different results in settings where AI is more or less likely to change daily work.

At the exposed end sat finance and insurance. At the other end were agriculture and forestry, with the least exposure to AI. Forestry and construction were the industries the release singled out as expected to face fewer future AI changes, and in those the effects were weaker. The release does not publish a separate number for each of the 18 industries, so the comparison is described only in terms of the two ends of the ranking.

What the finance and insurance result does and does not say

The finding that gives the study its sharpest edge concerns the most exposed sectors. According to the original WSU release dated October 12, the effects were stronger in finance and insurance, and the EurekAlert copy words it as the relationship being magnified in industries such as insurance and finance that are more heavily exposed to AI disruption. The attribution matters: finance and insurance are named together, as examples of heavily exposed industries, and neither text assigns the result to finance alone or provides a ranking among the 18 industries beyond the exposure scale.

Read with care, the result describes where the link between insecurity and justified misbehavior was more pronounced, not where employees are most likely to be caught slacking. A stronger association in a sector is compatible with low absolute rates of misconduct there, and the release gives no rates. It would be a stretch to say that worrying about AI makes a given banker or claims adjuster cut corners.

Because the data come from a survey, the direction of influence cannot be fixed from the numbers alone. Workers who already feel disengaged could be more alarmed about automation, or a third factor such as a poor relationship with a manager could drive both feelings. The authors’ framework treats insecurity as the starting point and the release describes it that way, but the design of a survey is what limits how far that framing can be pushed. The authors declare no conflicts of interest.

What employers can take from the findings

Probst’s own framing points toward communication. If insecurity about an occupation’s future is the starting point, then organizations that say plainly how they plan to use AI, and what that means for particular jobs, are addressing the first link in the chain rather than punishing behavior that appears further down it. The release offers that as the researchers’ reading, not as a tested intervention, and no experiment in the paper checks whether such reassurance reduces the behaviors.

For readers following the AI-and-work debate, the study is a reminder that the effects of automation are not limited to payroll numbers. Anxiety about replacement can shape how people behave in jobs they still hold. Journal details are listed on the publisher’s Occupational Health Science page, and the open question is how large these effects are outside the survey sample and whether they hold up when workers are followed over time.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

Make AI work for you: Get The Neuron →

We may earn a commission if you sign up.