Employees who regularly use artificial intelligence at work were more likely to report psychological stress than comparable non-users, according to a newly published Finnish study examining the relationship between workplace automation and employee well-being.
The research published November 13 in the British Journal of Industrial Relations found that regular AI use was associated with a 25.6% higher prevalence of reported stress.
The findings challenge the assumption that automating professional tasks necessarily reduces workplace pressure.
However, the researchers emphasized that the study identifies an association, not evidence that AI directly causes stress.
Researchers Analyze More Than 3,600 Workers
University of Turku researchers Thang Dang and Phuong Huynh analyzed data from the 2023 and 2024 Finnish Working Life Barometers.
Their study included 3,602 workers, comprising 552 regular AI users and 3,050 respondents who used AI less frequently or not at all.
Regular users were defined as employees using AI tools daily or weekly, including chatbots, virtual assistants, machine translation, speech recognition and other machine-learning applications.
The researchers applied propensity-score matching to compare employees with similar observable characteristics.
They found that regular AI users had a 4.4-percentage-point higher probability of reporting stress than matched non-users.
With 17.2% of workers in the full sample reporting stress, the estimated difference represented a 25.6% relative increase.
The researchers cautioned that their cross-sectional approach cannot rule out reverse causality or unmeasured characteristics that influence both AI adoption and stress.
Longer Working Hours May Explain the Relationship
The researchers examined two possible explanations: heavier work demands and anxiety about job security.
Their results favored the first explanation.
Regular AI use was associated with a 6.3-percentage-point higher probability of unpaid work outside normal working hours.
The association with intense work absorption was positive but less statistically certain.
Meanwhile, the relationship with perceived job insecurity was not statistically significant.
These findings suggest that the psychological consequences of AI adoption may involve how technology changes workloads rather than simply concerns about replacement.
The researchers also observed stronger individual statistical associations among men, non-supervisors, union members and employees in medium-sized workplaces.
However, differences between these groups were not statistically significant, preventing firm conclusions about which employees face greater risks.
European Research Shows Wider Workplace Pressures
A separate Eurofound report on algorithmic management found that 19% of EU employees reported computer-based performance monitoring, while 18% experienced automated task allocation and 10% had their working hours determined by software.
Employees subject to algorithmic management generally reported higher job demands and greater job strain.
These systems are not identical to the AI tools examined in Finland, but the findings offer broader context about digitally mediated work.
Meanwhile, a November 17 OECD and Eurofound working paper found that information-processing and problem-solving tasks had become more prominent in work organization between approximately 2011 and 2024.
The authors cautioned that rising educational attainment explained some changes more strongly than AI exposure.
What the Findings Mean for AI Companies
For founders and product leaders, the Finnish research raises questions about how workplace AI success should be measured.
Reducing the time required for individual tasks may not improve employee well-being if organizations simultaneously increase output expectations or extend working hours.
This creates a product-design and management challenge: automation should be evaluated alongside workload, autonomy, after-hours responsibilities and employee feedback.
The researchers called for longitudinal studies to establish whether AI adoption contributes to changes in mental health over time.
Until stronger causal evidence emerges, the findings provide a reason to examine workplace AI beyond productivity statistics.
The practical test is not simply whether employees complete more tasks, but whether automation improves how work is organized and experienced.