Flexible working boosts productivity most strongly (β=0.562), and human-centered AI adoption yields a smaller but significant productivity gain (β=0.263); digital leadership alone shows no measurable direct benefit.

Main Finding

Human-centric AI adoption and work flexibility both positively affect employee productivity in Industry 5.0 contexts, with work flexibility the strongest predictor (β = 0.562, p < 0.001). Human-centric AI has a significant positive effect (β = 0.263, p = 0.028). Digital leadership showed no statistically significant direct effect on productivity (β = −0.094, p = 0.275), suggesting leadership matters mainly when backed by appropriate technology, flexibility, and supporting resources.

Key Points

• The study integrates Human‑Centric AI Adoption, Digital Leadership, and Work Flexibility into a single model to explain employee productivity in the digital era (Industry 5.0).
• Theoretical framing: Resource‑Based View (RBV) and Dynamic Capabilities Theory (DCT). AI and flexibility are treated as strategic resources; digital leadership as a dynamic capability to coordinate them.
• Empirical findings:
  • Work Flexibility → Employee Productivity: strongest positive effect (β = 0.562, p < 0.001).
  • Human‑Centric AI Adoption → Employee Productivity: positive and significant (β = 0.263, p = 0.028).
  • Digital Leadership → Employee Productivity: not significant (β = −0.094, p = 0.275).
• Interpretation: Flexible work arrangements and AI tools designed to augment rather than replace workers boost perceived productivity. Digital leadership alone may not increase productivity unless technological infrastructure, employee capabilities, and flexible arrangements are in place (consistent with prior work showing leadership effects can be indirect/mediated).
• Limitations noted by the author: cross‑sectional survey design, reliance on self‑reported productivity measures, and limited attention to potential mediators or moderators (e.g., technical self‑efficacy, organizational readiness). The paper targeted ~300–500 responses from diverse sectors (manufacturing, IT, education, administrative services, public institutions), but the exact final sample size is not reported in the excerpts provided.

Data & Methods

• Research design: quantitative, cross‑sectional survey.
• Population/sample: employees from multiple sectors undergoing digital transformation (manufacturing, IT, education, admin services, public institutions); questionnaire distributed online. Targeted ~300–500 valid responses; participation voluntary and anonymous.
• Measures: validated/adapted scales
  • Human‑Centric AI Adoption: AI task support, employee–AI collaboration, skill enhancement (adapted from Shchepkina et al., 2024).
  • Digital Leadership: vision, support for innovation, digital communication, knowledge sharing (Shin et al., 2023).
  • Work Flexibility: remote work, scheduling autonomy, work–life balance (Mazur & Chukhray, 2023; Holovchenko, 2024).
  • Employee Productivity: perceived efficiency, work quality, task completion (Elsawy & Abu‑Alhaija, 2026; Williams & Anderson, 1991; Podsakoff et al., 1990).
• Analysis: Partial Least Squares Structural Equation Modeling (PLS‑SEM) via SmartPLS 4. Measurement model checks (factor loadings, Cronbach’s alpha, composite reliability, AVE), discriminant validity (Fornell‑Larcker, cross‑loadings). Structural model evaluated with path coefficients, t‑values, p‑values, f², Q², and R².

Implications for AI Economics

• Complementarity and productivity accounting:
  • The results support the view that AI is productivity‑complementary to labor when designed and deployed in a human‑centric way. Economic models of AI adoption should incorporate complementarities between AI and workplace practices (especially flexible work) rather than treating AI as a pure labor substitute.
  • Empirical productivity decompositions (firm or macro) should consider the joint contribution of organizational practices (flexibility) and AI adoption to measured output per worker.
• Labor demand and skill upgrading:
  • Positive productivity effects from human‑centric AI imply potential for task reallocation toward higher‑value activities; economists should model how demand shifts across skill bundles and the resulting wage and employment dynamics.
  • Policies that facilitate digital upskilling and reduce technostress (training, digital literacy, change management) will likely raise the realized productivity gains from AI.
• Role of management and institutions:
  • Digital leadership alone is insufficient to raise productivity — economists and policymakers should treat managerial capabilities as necessary but not sufficient conditions; they must be paired with investments in technology, flexible work design, and employee support.
  • When modeling diffusion and returns to AI, include interactions between leadership quality, technology readiness, and flexible workplace arrangements.
• Policy and evaluation recommendations:
  • Promote human‑centric AI standards, funding for worker training, and incentives for flexible work infrastructure to maximize productivity gains from AI.
  • Encourage longitudinal and causal studies (e.g., panel data, difference‑in‑differences, randomized rollouts) to estimate causal productivity effects and persistence over time.
• Research agenda for AI economics:
  • Quantify the macroeconomic impact of combined AI + flexibility adoption on aggregate productivity, labor shares, and distributional outcomes.
  • Investigate heterogeneous effects by sector, firm size, and task composition (routine vs. non‑routine).
  • Estimate the degree to which managerial capabilities moderate AI returns, and the costs/benefits of investments in complementary human capital and organizational practices.

Concise managerial takeaway: prioritize flexible work policies and human‑centric AI designs, and pair digital leadership development with concrete investments in infrastructure and employee support to capture productivity gains.