The research team surveyed 499 information technology employees working in AI-enabled organizations in Hyderabad, India, one of the country’s most dynamic technology hubs. Using a cross-sectional survey design, the researchers measured how regularly employees used AI tools in their daily tasks, how much stress those technologies generated, how much they trusted AI systems to perform work-related duties reliably, and how psychologically well they felt while working. All constructs were captured on seven-point Likert scales, with items adapted from established measurement traditions, including Davis’s foundational work on technology acceptance, Tarafdar and colleagues’ technostress framework, McKnight’s trust-in-technology scales, and Zheng and colleagues’ well-being measures.

To make sense of the data, the team turned to three complementary theoretical pillars. The Job Demands–Resources model provided the scaffolding for understanding how workplace characteristics can act as either demands that deplete employees or resources that energize them. The Unified Theory of Acceptance and Use of Technology 2, known as UTAUT2, helped explain the motivational drivers behind AI adoption itself. Social Exchange Theory rounded out the framework by framing the relationship between employees and the intelligent systems they work alongside as a kind of reciprocal exchange, in which trust functions as the currency of cooperation.

Methodologically, the study was rigorous by the standards of organizational psychology. The researchers first ran exploratory factor analysis to uncover the underlying structure of their measures, then confirmatory factor analysis to verify that each construct held together as intended. Finally, they employed covariance-based structural equation modeling, using IBM SPSS 29 and AMOS 28, to test the full network of hypothesized relationships simultaneously. Bootstrapping procedures were used to assess mediation, a technique that repeatedly resamples the data to estimate how confidently an indirect effect can be claimed. The study protocol was reviewed and approved by an Institutional Ethics Committee, with formal exemption approval granted on 6 May 2026, and all participants provided electronic informed consent before completing the anonymous online questionnaire.

The headline result is deceptively simple: AI adoption has a positive direct impact on employees’ psychological well-being. Workers who regularly use AI tools to perform their tasks report feeling psychologically better at work, a finding consistent with the growing body of evidence that automation and augmentation can relieve employees of drudgery, sharpen decision-making, and make work feel more meaningful. But the study did not stop at that direct effect. It also found that AI adoption increases trust in AI and, at the same time, increases technostress, the technology-induced strain that arises when systems demand constant learning, accelerate workloads, blur boundaries, and create fears of obsolescence.

Here is where the story becomes more complicated and more interesting. Technostress, measured with items such as whether using AI technologies increases one’s workload, negatively impacts employees’ psychological well-being. Trust in AI, captured by items asking whether employees consider AI systems reliable in performing work-related tasks, positively impacts well-being. In other words, AI adoption sends two opposing signals through the workforce simultaneously. One pathway transmits the negative effects of new technology through heightened stress, while the other transmits positive effects through the confidence employees place in the systems they use. The net experience of any given employee is the sum of these competing currents.

The most consequential finding, however, concerns mediation. The researchers hypothesized that technostress would negatively mediate the relationship between AI adoption and well-being, transmitting harm through increased technology-related stress, while trust in AI would positively mediate the same relationship by fostering confidence in AI systems that is itself associated with psychological well-being. When the bootstrapping analysis was complete, the results showed something even stronger than partial mediation: technostress and trust in AI fully mediate the relationship between AI adoption and employee well-being. In statistical terms, this means that AI adoption does not influence well-being through some mysterious direct residue once these two mechanisms are accounted for. Its entire psychological footprint travels through the twin channels of stress and trust.

That full-mediation result carries a practical punch for organizations racing to embed AI into their operations. If the well-being consequences of AI adoption flow entirely through technostress and trust, then the levers that matter most are not the technologies themselves but the human conditions surrounding them. Companies that roll out AI without managing workload pressures, skill anxiety, and constant connectivity may find that the stress pathway dominates, eroding the very well-being that productivity gains were supposed to enhance. Conversely, organizations that invest in transparency, reliability, and demonstrable competence of their AI systems can strengthen the trust pathway, converting adoption into a genuine psychological resource. The study’s authors frame this as a call for human-centered AI practices and responsible AI governance within organizations, positioning employee psychology as a central concern of AI strategy rather than an afterthought.

The research also fills a notable gap in the literature. While past studies have demonstrated positive effects of AI on organizational productivity and efficiency, the impact of AI adoption on employees’ psychological well-being has remained unclear, with most work examining either the positive or the negative pathway in isolation. By modeling both the dark side and the bright side of AI adoption within a single study, the researchers provide a more complete map of how intelligent technology reshapes the inner lives of workers. The finding that AI adoption simultaneously raises technostress and builds trust captures the ambivalence many employees report feeling: energized by capable tools, yet unsettled by the pace of change they impose.

Certain limitations are worth keeping in view. The study is cross-sectional, meaning it captures a single moment in time and cannot definitively establish causal direction; it is theoretically possible, for example, that employees who feel psychologically well are simply more inclined to adopt and trust AI. The sample was drawn from IT employees in a single Indian city, a population that is both unusually familiar with technology and embedded in a rapidly transforming labor market, so generalizing to other industries, cultures, or occupational groups requires caution. The authors also note that the article was shared early as a peer-reviewed, accepted version subject to further edits before the final Version of Record. Still, the core message stands out with unusual clarity for workplace research: the psychological consequences of AI are not written into the algorithms. They are written into the stress employees feel and the trust they extend, and both of those are things organizations can actively shape.

Subject of Research: The mediating roles of technostress and trust in AI in the relationship between workplace AI adoption and employee psychological well-being

Article Title: Technostress and trust in AI mediate the association between AI adoption and employee well-being

Article References: Prasad, K. D. V., Srinivas, V., Singh, S., Kothari, H., Nag, D., & Pathak, A. (2026). Technostress and trust in AI mediate the association between AI adoption and employee well-being. Discover Psychology. https://doi.org/10.1007/s44202-026-00928-9

Image Credits: AI Generated

DOI: 10.1007/s44202-026-00928-9

Keywords: AI adoption, technostress, trust in AI, employee well-being, Job Demands-Resources model, UTAUT2, Social Exchange Theory, structural equation modeling, organizational psychology, IT employees, human-centered AI, responsible AI governance