Truace
Subscribe
60/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

Workers who use AI more learn more, but fear of job loss cuts it, a 420-person study found

Study title: Opportunity and Threat in Parallel: How Artificial Intelligence Application Predicts Employees’ Proactive Learning Behavior Through Parallel Primary Appraisals

Evidence type
Observational study
Sample
n = 420
Sources
1 sourcePeer-reviewed
Published
3 min read
Truvace record
TRV-2026-2520cite or check it

Study title: Opportunity and Threat in Parallel: How Artificial Intelligence Application Predicts Employees’ Proactive Learning Behavior Through Parallel Primary Appraisals

A four-wave panel study of 420 employees in eight industries in China, surveyed monthly from March to June 2025, found that more AI use at work was associated with more self-initiated learning of new skills. After adjusting for each person's earlier learning level, the total association was 0.235.

Read the study abstract

As organizations embed artificial intelligence (AI) in everyday work, understanding how employees respond has become a pressing question for human resource management. Drawing on the cognitive appraisal theory of stress, this study examines how AI application relates to employees' proactive learning behavior through two parallel primary appraisals: opportunity perception and threat perception. It also tests proactive personality both as a direct predictor and as a moderator at two theoretically distinct location…

56Gain side

How the two sides scoreClose call

54Problem side

The gain

In a four-wave study of 420 Chinese employees, heavier AI use at work was linked to more self-initiated skill learning (total association 0.235), mainly among workers who saw AI as an opportunity.

The problem

In the same study, employees who saw AI as a threat to their jobs did less proactive learning (indirect association -0.140), though this pathway weakened under the strictest statistical adjustment.

A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.

What the scores mean

G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.

P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.

In brief

That net figure combines two opposing paths. Employees who saw AI as an opportunity learned more (indirect association 0.193), while those who saw it as a threat to their jobs learned less (−0.140). The two views were statistically independent, and 18.8% of respondents scored above the midpoint on both.

The threat path was less solid than the opportunity path. It disappeared under the most stringent adjustment, which treated threat perception as change from its earlier level, but was stronger over a three-month window (−0.164 indirect). The authors describe the findings as association, not cause.

Main points

  1. The sample was 420 matched employees (from 510 responses) across services, manufacturing, public institutions, high technology, construction, education and government; 53.1% were frontline staff and 46.9% managers.
  2. Opportunity perception predicted more proactive learning (β = 0.476) and threat perception predicted less (β = −0.319) in the primary baseline-adjusted model.
  3. Without baseline adjustment the total association was larger (0.328); adjustment cut the opportunity path by roughly 30% and left the threat path essentially unchanged.
  4. Under full autoregressive adjustment the opportunity indirect path stayed significant but small (0.046), while the threat indirect path was not significant (−0.003).
  5. Among respondents, 18.8% scored above the scale midpoint on both opportunity and threat, 22.9% on opportunity only, 24.8% on threat only and 33.6% on neither.
  6. Proactive personality shaped how AI was read: the AI-to-opportunity slope rose from 0.144 to 0.727 from low to high proactive personality, and the AI-to-threat slope fell from 0.711 to 0.105.
  7. Proactive personality did not change how strongly a formed appraisal turned into learning (second-stage interactions not significant).
  8. The threat-to-learning path ranged from −0.035 in construction to −0.529 in public institutions, in subsamples of 36 to 104 respondents.

The full story

Participants were recruited from organizations where AI is used in daily tasks. AI use was measured in March 2025, baseline learning in April, opportunity and threat perception in May, and proactive learning in June. Attention checks, response-time filters and straight-lining screens removed 90 responses, and retained and excluded participants did not differ on the main Wave 1 variables.

The primary model adjusted the June learning outcome for its April baseline. AI application predicted opportunity perception (β = 0.406) and threat perception (β = 0.439). The decomposition was exact: 0.193 − 0.140 + 0.182 = 0.235.

The researchers say the opportunity pathway survived every specification, while the threat pathway depended on how baseline adjustment was done. AI use predicted the level of threat perception but not month-to-month change in it. Threat perception was highly stable over time (r = 0.858).

Proactive personality was positively linked to opportunity perception (β = 0.202), negatively to threat perception (β = −0.209) and positively to learning (β = 0.127). Employees who already learned proactively also read AI more favorably, so the relationship may run in both directions.

The authors suggest employers frame AI as a complement to workers during onboarding and pair benefit messaging with reskilling and job-scope clarity, but say their design did not test any intervention.

What this doesn’t fix

All measures were employee self-reports, the design cannot establish causation or direction of influence, the sample came from a single cultural context, and the AI measure could not separate generative AI from other AI types.

Sources

1 on file
  1. Behavioral Sciencesdoi:10.3390/bs16091713
    Peer-reviewed2026-09-21

Talk it over

Four questions to start onThis record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?

Start with a question

Workers who use AI more learn more, but fear of job loss cuts it, a 420-person study foundhttps://truvace.com/item/opportunity-and-threat-in-parallel-how-artificial-intelligence-application-predi

  1. This record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?

    The gain In a four-wave study of 420 Chinese employees, heavier AI use at work was linked to more self-initiated skill learning (total association 0.235), mainly among workers who saw AI as an opportunity.

    The problem In the same study, employees who saw AI as a threat to their jobs did less proactive learning (indirect association -0.140), though this pathway weakened under the strictest statistical adjustment.

    Answer in the thread

  2. How do we know? What can this kind of evidence show, and what can’t it?

    The evidence This rests on an observational study with a sample of 420, from Behavioral Sciences.

    Answer in the thread

  3. What does it leave open? What else would you want to know before acting on it?

    The record’s own caveat All measures were employee self-reports, the design cannot establish causation or direction of influence, the sample came from a single cultural context, and the AI measure could not separate generative AI from other AI types.

    Answer in the thread

  4. Have you seen AI used this way where you study, work or live? Did it go the same way?

    Answer in the thread

The questions are the same on every record. The details under them are taken from this record as written.

What readers say

More from Labor