TRV-2026-2914Version 1 · Published

Saved · see the current record

Why this version exists

Certified into the record

The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2914
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-14T12:02:22.352410Z
status: published
lens: p_space
sector: labor
headline: AI and Stress
dek: ABSTRACT Given the spread of artificial intelligence (AI) tools at work, understanding their health implications is critically important. We use rich survey data from Finland and apply propensity score matching that accounts for observable differences between regular AI users and non‐users to estimate the relationship between reported AI usage at work and workers’ reported stress. Regular AI use is associated with a 4.4 percentage point higher probability of reporting stress, equivalent to a 25.6% increase relat…
gain_title: (none)
problem_title: A survey study of Finnish workers found that regular AI use at work was associated with a 4.4 percentage point higher probability of reporting stress, a 25.6% increase over baseline.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: A survey study of Finnish workers found that regular AI use at work was associated with a 4.4 percentage point higher probability of reporting stress, a 25.6% increase over baseline.
problem_evidence: Regular AI use is associated with a 4.4 percentage point higher probability of reporting stress, equivalent to a 25.6% increase relative to baseline.
quick_read: A study using survey data from Finland found that workers who regularly use AI at work were 4.4 percentage points more likely to report stress than comparable non-users. That equals a 25.6% increase relative to the baseline rate.

The authors matched regular AI users with non-users on observable characteristics using propensity score matching. They then tested two explanations for the pattern: worry about losing a job, and heavier work demands.

The results point toward heavier demands. AI use was associated with more unpaid work outside working hours and, less precisely, with intense work absorption, but not with job insecurity. Subgroup differences were not statistically significant.
limitation: The study reports an association based on survey data, and the subgroup estimates are exploratory: differences across subgroups were not statistically significant.
tag: Evidence-backed problem
key_points: Regular AI use was associated with a 4.4 percentage point higher probability of reporting stress among Finnish workers. | That gap equals a 25.6% increase relative to the baseline stress rate. | The study used propensity score matching to account for observable differences between regular AI users and non-users. | AI use was associated with more unpaid work outside working hours. | AI use was linked, less precisely, to intense work absorption. | AI use was not associated with job insecurity, which weakens the job-loss explanation. | Exploratory estimates were larger for men, non-supervisors, unionized workers and employees in medium-sized workplaces.
rundown: The research draws on rich survey data from Finland and compares workers who report regularly using AI tools with those who do not. Both AI use and stress are self-reported, and propensity score matching adjusts for differences that can be observed between the two groups.

The authors weighed two competing explanations for the link. One is that AI raises fears about job loss; the other is that it raises work demands. The data fit the second better, since AI use tracked with unpaid work outside working hours and, less precisely, with intense work absorption, while job insecurity showed no association.

Exploratory subgroup estimates were larger and individually significant for men, non-supervisors, unionized workers and employees in medium-sized workplaces. The differences across subgroups were not statistically significant, so they should not be read as confirmed gaps.
sources:
- peer_reviewed | British Journal of Industrial Relations | https://doi.org/10.1111/bjir.70084 | 2026-11-13
prev: 0000000000000000000000000000000000000000000000000000000000000000
SHA-256 hash
c72f4684c8cdf0ab350ce65cc7fd663cb9de2cb83c27539e1446b1415c7a4e27
Previous version’s hash
0000000000000000000000000000000000000000000000000000000000000000
Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-2914 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.