0 cumulative citations
View corpus contextSurveyed Chinese employees who collaborate with AI report higher creativity, and that boost appears to flow through stress-related pathways: demanding but growth-oriented AI tasks (challenge stressors) raise creative output — and, contrary to expectations, so do hindrance-type demands — while organizational support amplifies the benefit of challenge stressors.
Human-AI Collaboration and Employee Creativity: A Dual-Pathway Perspective of Challenge and Hindrance Stressors
openalex
correlational
low evidence
7/10 relevance
Full text
usable extracted full text
DOI
Source
PDF
In a cross-sectional survey of 292 Chinese employees, human–AI collaboration is positively associated with employee creativity, with both challenge and (unexpectedly) hindrance stressors mediating that relationship and perceived organizational support strengthening the challenge-stressors → creativity link.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
Artificial intelligence (AI) is now widely used in the workplace. As a result, human–AI collaboration plays an increasingly important role in employee creativity. However, the psychological mechanisms underlying this relationship remain insufficiently understood. Drawing on the Challenge–Hindrance Stressor Framework and Organizational Support Theory, this study examines the mediating roles of challenge and hindrance stressors in the relationship between human–AI collaboration and employee creativity. It also investigates the moderating role of perceived organizational support. Survey data were collected from 292 Chinese employees with human–AI collaboration experience and analyzed using SPSS 27.0, AMOS 24.0, and the PROCESS macro. The results show that challenge stressors significantly mediate the positive relationship between human–AI collaboration and employee creativity. Hindrance stressors also exhibit a significant indirect effect; however, the effect is positive, contrary to the hypothesized negative direction. In addition, perceived organizational support strengthens the positive relationship between challenge stressors and employee creativity. Overall, these findings further clarify the psychological mechanisms through which human–AI collaboration influences employee creativity and highlight the important role of organizational support in strengthening the positive effect of challenge stressors on Employee creativity.
Summary
Main Finding
Human–AI collaboration is positively associated with employee creativity. This relationship operates through a dual-stressor pathway: challenge stressors significantly mediate the positive effect, and—unexpectedly—hindrance stressors also show a positive indirect association with creativity. Perceived organizational support (POS) strengthens the positive effect of challenge stressors on creativity but does not buffer the effect of hindrance stressors.
Key Points
- Theoretical framing: combines the Challenge–Hindrance Stressor Framework with Organizational Support Theory to explain how human–AI collaboration affects creativity via stress processes and organizational resources.
- Human–AI collaboration provides knowledge, automation of routine tasks, and quick feedback—resources that can boost creativity (Hypothesis 1 supported; β = 0.425, p < 0.001).
- Dual-pathway stressors:
- Challenge stressors (learning, growth opportunities) rise with human–AI collaboration and positively predict creativity (β = 0.316, p < 0.001); they mediate the AI→creativity link.
- Hindrance stressors (complexity, adaptation burden) also increase with human–AI collaboration, but contrary to expectations they were positively associated with creativity in this sample (β = 0.248, p < 0.001) and produced a positive indirect effect.
- Moderation by POS:
- POS significantly strengthens the positive relationship between challenge stressors and creativity (interaction β = 0.096, p < 0.05).
- POS did not significantly moderate the hindrance stressor → creativity relationship (interaction β = 0.018, ns).
- Measurement & validity: constructs show acceptable-to-excellent internal consistency (Cronbach’s α .789–.894) and a five-factor CFA fit the data well (χ²/df = 1.082; CFI = 0.992; RMSEA = 0.017).
Data & Methods
- Sample: N = 292 Chinese employees with human–AI collaboration experience (collected via Wenjuanxing, May 17–31, 2026). Balanced gender (51.0% male), wide age distribution (largest 30–40 years old), diverse industries (financial services, manufacturing, services).
- Design: Cross-sectional, self-report online survey.
- Measures:
- Human–AI Collaboration (Kong et al., 2023), α = .894
- Challenge Stressors (adapted Cavanaugh et al., 2000), α = .874
- Hindrance Stressors (Cavanaugh et al., 2000), α = .859
- Perceived Organizational Support (Eisenberger et al., 1997), α = .883
- Employee Creativity (Farmer et al., 2003), α = .789
- Controls: age, gender, education, organizational tenure, job position, employment status, job function, industry.
- Analyses: SPSS 27.0, AMOS 24.0, PROCESS macro for mediation/moderation. Confirmatory factor analysis supported discriminant validity. Harman’s single-factor test and VIFs indicated limited common-method bias and multicollinearity.
- Key reported statistics:
- Correlations: Human–AI collaboration with challenge stressors r = 0.420, hindrance r = 0.253, creativity r = 0.449** (p < 0.001).
- Hierarchical regressions: H1 supported (β = 0.425, p < 0.001); H3a supported (challenge → creativity β = 0.316, p < 0.001); H3b not supported (hindrance → creativity positive β = 0.248, p < 0.001). Moderation: challenge×POS significant; hindrance×POS not significant.
- Limitations of methods: cross-sectional self-report design, single-country sample (China), potential endogeneity or reverse causality not addressed.
Implications for AI Economics
- Productivity vs. adaptation costs: Human–AI collaboration raises productive inputs (data, automation, cognitive augmentation) that increase creative outputs, supporting models where AI complements human creativity and raises firm-level innovation capacity.
- Human capital and training investments: Because human–AI collaboration elevates both challenge and hindrance demands, firms and policymakers should account not only for direct productivity gains but also for up-front and ongoing investments in training, onboarding, and support systems. POS-like organizational investments (supervision, recognition, resource access) amplify the creativity gains from challenge-type demands and are economically valuable.
- Returns to complementary investments: Results suggest positive returns to investing in organizational support (e.g., mentoring, psychological safety, recognition) to unlock the creative potential of AI augmentation—this can be translated into firm-level decisions on HR spending and into macro-level assessments of AI adoption spillovers.
- Labor-market implications: The coexistence of higher challenge and hindrance stressors highlights heterogeneous effects on workers—some workers convert disruption into skill accumulation and higher creative productivity, while others may face burdens. Economic models should allow for heterogeneity in adjustment costs and creative output responses to AI exposure.
- Measurement and accounting of AI impact: Standard productivity metrics may undercount innovation-related gains from AI if they ignore stressor-mediated pathways and organizational support. Empirical estimates of AI’s effect on innovation should control for organizational support and account for both positive and unexpected associations (e.g., hindrance stressors sometimes correlating positively with creativity).
- Policy and organizational design: Promote policies and firm practices that increase perceived organizational support (training subsidies, mental health resources, recognition systems) to maximize creative gains from AI. Given unexpected positive links between hindrance stressors and creativity here, policymakers should fund research to identify boundary conditions (e.g., industry, task type) before generalizing.
- Research priorities for AI economics:
- Longitudinal and quasi-experimental studies to estimate causal effects of human–AI collaboration on innovation and wages.
- Heterogeneity analysis by occupation, task routineness, and worker skill to model differential dynamic adjustment costs.
- Cost–benefit analyses of organizational support interventions to quantify their impact on firm-level innovation returns.
(Article: Du, Y., Zhang, C.*, & Gong, L. (2026). Human–AI Collaboration and Employee Creativity: A Dual-Pathway Perspective of Challenge and Hindrance Stressors. Asia Pacific Economic and Management Review, 3(5). DOI: https://doi.org/10.62177/apemr.v3i5.1653)
Assessment
Paper Typecorrelational
Evidence Strengthlow — All data are cross-sectional and self-reported from a single survey; mediation was tested using regression/PROCESS but temporal precedence and unobserved confounding are not addressed, so causal interpretation is weak and reverse causality / common-method/artifact explanations remain plausible.
Methods Rigormedium — The authors used validated scales with reported Cronbach's alphas, conducted CFA to check factor structure, reported correlations, checked collinearity, and used hierarchical regression and interaction tests; however, reliance on a single-source cross-sectional design, limited common-method controls (Harman's single-factor test only), and no robustness checks (e.g., alternative specifications, instrumental variables, longitudinal data) reduce rigor.
SampleOnline convenience sample of 292 Chinese employees with human–AI collaboration experience collected via Wenjuanxing between May 17–31, 2026; balanced gender (51% male), majority aged 30–40, mixed education levels, range of job functions and industries (largest: financial services, manufacturing, service), both regular and non-regular employees, self-reported measures for all constructs.
Themeshuman_ai_collab innovation org_design skills_training
IdentificationCross-sectional observational survey using self-reported measures; causal claims are tested via mediation and moderation analyses (hierarchical regression and PROCESS macro) but identification relies on assumed temporal ordering and no omitted variable bias rather than any experimental or quasi-experimental design.
GeneralizabilitySingle-country (China) sample may limit transferability to other cultural and institutional contexts, Online, self-selected respondents (Wenjuanxing) introduce selection bias and may not represent broader workforce, Cross-sectional self-report measures limit causal generalization to real productivity or firm-level outcomes, Sample includes only employees with some human–AI experience; findings may not apply to workers without AI exposure, Industry composition (over-representation of financial services, manufacturing, services) may bias applicability to other sectors
Claims (8)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| Human–AI collaboration positively influences employee creativity. Creativity | positive | Employee creativity, defined as generating novel and useful ideas related to methods, processes, products, or services. |
Reading fidelity
high
Study strength
medium
|
n=292
β = 0.425, p < 0.001; ΔR² = 0.175, p < 0.001
|
| Challenge stressors are positively associated with employee creativity. Creativity | positive | Employee creativity. |
Reading fidelity
high
Study strength
medium
|
n=292
β = 0.316, p < 0.001
|
| Hindrance stressors are positively associated with employee creativity in this sample, contrary to the hypothesized negative relationship. Creativity | positive | Employee creativity. |
Reading fidelity
high
Study strength
medium
|
n=292
β = 0.248, p < 0.001
|
| Perceived organizational support strengthens the positive relationship between challenge stressors and employee creativity. Creativity | positive | Employee creativity as a function of challenge stressors and perceived organizational support. |
Reading fidelity
high
Study strength
medium
|
n=292
β = 0.096, p < 0.05
|
| There is no statistically significant evidence that perceived organizational support moderates the relationship between hindrance stressors and employee creativity. Creativity | null_result | Employee creativity as a function of hindrance stressors and perceived organizational support. |
Reading fidelity
high
Study strength
medium
|
n=292
β = 0.018, p > 0.05
|
| Challenge stressors significantly mediate the positive relationship between human–AI collaboration and employee creativity. Creativity | positive | Employee creativity through challenge stressors as a mediator of human–AI collaboration. |
Reading fidelity
high
Study strength
low
|
n=292
|
| Hindrance stressors have a significant positive indirect effect in the relationship between human–AI collaboration and employee creativity, contrary to the hypothesized negative indirect effect. Creativity | positive | Employee creativity through hindrance stressors as a mediator of human–AI collaboration. |
Reading fidelity
high
Study strength
low
|
n=292
|
| Human–AI collaboration is positively correlated with challenge stressors, hindrance stressors, employee creativity, and perceived organizational support. Creativity | positive | Challenge stressors, hindrance stressors, employee creativity, and perceived organizational support. |
Reading fidelity
high
Study strength
low
|
n=292
r = 0.420, 0.253, 0.449, and 0.318 respectively; all p < 0.001
|