Fear of AI Job Loss Drives Workers Toward Counterproductive Behavior, Study Finds

A sweeping new study led by researchers at Washington State University has found that workers who fear artificial intelligence could render their entire occupation obsolete are significantly more likely to engage in counterproductive workplace behaviors, from slacking off and surfing the web to coming in late and even taking a co-worker’s lunch from the break room refrigerator. The research, published in the journal Occupational Health Science, offers one of the most detailed portraits yet of how the psychological shadow cast by AI adoption reshapes everyday conduct on the job, and it suggests that the human costs of automation anxiety may be far greater than employers currently appreciate.

The study was led by Tahira M. Probst, a professor of psychology at Washington State University Vancouver, working with graduate student Rebecca Lindgren and collaborators in Italy and Belgium. Rather than focusing narrowly on whether individual employees worry about losing their current position, the team investigated a deeper and potentially more corrosive form of insecurity: the fear that one’s whole occupation might disappear. The distinction matters, the researchers argue, because a worker who loses a job in a stable profession can typically find equivalent work elsewhere, whereas a worker whose occupation is being automated away faces the prospect that no comparable role will exist at any employer.

Probst illustrated the difference with a concrete comparison. If a bartender’s employer goes out of business, that bartender can seek work at another bar, because the profession itself is not going away. But for someone in a field such as medical transcription, where coding and transcription tasks can now be substantially automated, being let go may mean the occupation itself is vanishing, leaving no parallel job to move into. It is this occupation-level insecurity, the study finds, that most powerfully predicts moral disengagement and subsequent misbehavior at work.

To measure how exposed different sectors are to AI-driven disruption, the researchers drew on an existing objective metric, the AI Industry Exposure index, which quantifies the potential for artificial intelligence to displace tasks within a given industry. They then surveyed workers across 18 industries about their attitudes and behaviors, pairing self-reported occupation insecurity with the objective exposure scores of their sectors. Finance and insurance workers registered some of the highest levels of AI exposure in the analysis, while those in agriculture and forestry showed the least. This design allowed the team to separate the subjective experience of fearing obsolescence from the structural reality of how much disruption a worker’s industry actually faces.

The psychological mechanism at the center of the study is moral disengagement, a well-established construct in organizational psychology describing the process by which people justify conduct they would ordinarily recognize as wrong. As Probst explained, workers generally know what is right and wrong and what organizational policies permit, but moral disengagement allows them to reframe questionable behavior as acceptable. In practical terms, an employee who feels doomed by automation may come to see cutting corners, wasting organizational resources, or undermining colleagues not as violations but as reasonable responses to an unfair situation, or as necessary means of self-protection.

The data showed a clear chain of effects. Workers worried about the future of their occupation were more likely to morally disengage, and that moral disengagement in turn predicted higher levels of self-reported counterproductive workplace behaviors. These ranged from relatively minor transgressions, such as browsing the internet when they should be working or arriving late for a shift, to behaviors with more direct interpersonal costs, such as taking a co-worker’s lunch. The pattern held across the sample, but its strength varied systematically with industry-level AI exposure.

Specifically, the effects were magnified in industries expected to face heavy future AI disruption, such as insurance and finance, and were less pronounced in sectors like forestry and construction where fewer AI-driven changes are anticipated. This dose-response-like pattern strengthens the study’s central claim, because it suggests the observed behaviors are not simply a product of generic job anxiety but are tied to the concrete prospect of AI reshaping a worker’s field. The more real the threat of occupational obsolescence appears, the more employees appear to loosen the internal restraints that normally govern workplace conduct.

The consequences extend well beyond individual misconduct. As Probst noted, counterproductive workplace behaviors damage the interpersonal relationships within an organization. When employees fail to show up, the burden of their work falls on colleagues; when they act in ways that impair their own or others’ ability to complete tasks properly, or when they squander organizational resources, the harm accumulates across the whole enterprise. The costs, she emphasized, are not confined to the bottom line but extend to the effective functioning of the organization itself, eroding the cooperation and trust on which productive work depends.

The findings arrive against a backdrop of rapidly expanding AI adoption in the workplace. Research has already shown that as intelligent robotics, automated systems, and algorithmic management spread, workers frequently experience heightened job insecurity, and that the resulting stress is associated with burnout, reduced job satisfaction and engagement, and worse physical and mental health. The new study extends this literature by demonstrating that such insecurity does not merely diminish wellbeing; it can actively push employees toward behaviors that harm their employers and co-workers, creating a feedback loop in which the anticipation of technological displacement degrades the very workplaces workers still inhabit.

For employers, the message of the research is that how organizations discuss and implement AI matters as much as the technology itself. The authors argue that companies should pay close attention to the way they communicate about automation and to the support they provide employees during transitions, in order to keep workers engaged and supported rather than frightened and alienated. Where AI adoption is framed as a threat to entire occupations, the study suggests, employees may respond by psychologically withdrawing from workplace norms; where it is managed transparently and paired with genuine pathways for retraining and redeployment, the corrosive effects on behavior may be blunted. As AI continues to penetrate sectors from finance to transcription services, the study stands as a reminder that the success of workplace automation will be measured not only in efficiency gains but in the morale and conduct of the humans who remain.

Subject of Research: The relationship between AI-related occupational insecurity, moral disengagement, and counterproductive workplace behaviors across industries

Article Title: Workers who fear AI replacement show worse workplace behavior

Article References: Workers who fear AI replacement show worse workplace behavior. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: artificial intelligence, job insecurity, occupational health, moral disengagement, counterproductive work behavior, workplace psychology, automation, employee wellbeing, organizational behavior, AI industry exposure, Washington State University, Occupational Health Science