AI adoption in Chinese sports firms stabilises supplier relationships but destabilises customer ones, driven chiefly by talent acquisition rather than logistics gains. The net effects vary by firm type and profitability, implying AI returns depend heavily on firm context and workforce strategies.
Can Artificial Intelligence Enhance the Stability of Supply Chain Systems for Sports Enterprises? Insights from Systems Theory and Supply Chain Management Theory
Using DML on a 2012–2023 panel of 45 Chinese listed sports firms, the authors find AI adoption increases supplier-side stability but decreases customer-side stability, with talent attraction (AI-related human capital) the main mediating channel and logistics efficiency not mediating effects.
In the digital economy, the effective use of artificial intelligence (AI) is crucial for maintaining supply chain stability (SCS) in sports enterprises (SEs). Leveraging systems theory and supply chain management theory, we construct a dual machine learning model (DML) to empirically assess the impact of AI on the SCS of SE. This analysis is based on panel data from 45 Chinese listed SEs over the period 2012–2023. The results indicate that AI significantly enhances supplier stability but notably reduces customer stability in SE. Talent attraction emerges as the primary mechanism, while logistics efficiency fails to fulfill its anticipated role. The impact of AI on SCS in SE exhibits heterogeneity based on enterprise type and profitability status. Our findings offer valuable insights for harnessing the potential of AI and fostering its deeper integration into the supply chains of SE.
Summary
Main Finding
AI adoption in Chinese sports enterprises (SEs) (2012–2023) has an asymmetric effect on supply chain stability (SCS): it materially strengthens supplier stability but materially weakens customer stability. Talent attraction is identified as the chief mediating channel; expected gains through logistics efficiency are not realized. Effects vary by enterprise type and profitability.
Key Points
- Empirical approach: authors use a dual machine learning (DML) framework to estimate the impact of AI on SCS while flexibly controlling for confounders.
- Sample: panel of 45 Chinese listed sports enterprises over 2012–2023.
- Directional results:
- Positive, significant effect of AI on supplier-side stability.
- Negative, significant effect of AI on customer-side stability.
- Mechanisms:
- Talent attraction (AI-related human capital) is the primary mechanism linking AI to SCS outcomes.
- Logistics efficiency does not mediate the AI → SCS relationship as hypothesized.
- Heterogeneity: the magnitude and sign of AI’s effects differ by firm type and by firms’ profitability status.
Data & Methods
- Data: firm-level panel (45 listed SEs) across 12 years (2012–2023). Focus on the sports enterprise sector in China.
- Estimation strategy: dual machine learning (DML) to estimate causal effects while using machine learning to flexibly model high-dimensional controls/nuisance functions and reduce bias from confounding.
- Mediator and heterogeneity analysis: decomposition to identify channels (talent, logistics) and subgroup analyses by enterprise type and profitability.
- Strengths: causal-focused estimation with modern ML tools; multi-year panel covering pre- and post-AI diffusion years in the sector.
- Limitations to note: relatively small, single-sector and single-country sample (listed firms only); potential measurement/construct validity issues for AI use and SCS components; DML requires strong ignorability assumptions for causal interpretation.
Implications for AI Economics
- Policy and workforce:
- Talent accumulation is a key mechanism—policy and firm-level investments in AI-related human capital are crucial to realize supplier-side benefits.
- Labor-market and training policies should anticipate reallocation effects that may harm customer-facing stability if not managed.
- Managerial strategy:
- Firms should recognize trade-offs: AI can make supplier relationships more stable but may destabilize customer relationships unless deployment is designed to preserve customer engagement.
- Because logistics efficiency did not mediate benefits here, managers should not assume across-the-board operational gains from AI; complementary investments (process redesign, customer-facing training) may be necessary.
- Valuation and investment:
- Heterogeneous effects imply that the returns to AI investment depend on firm type and profitability—investors and executives should incorporate firm-specific context when forecasting AI payoffs.
- Research agenda:
- Broader samples (other sectors, private firms, other countries) to test external validity.
- Finer-grained measurement of AI use and of supplier vs customer stability dynamics.
- Experimental or quasi-experimental designs to strengthen causal claims about mechanisms (e.g., randomized talent-training interventions).
Assessment
Paper Typequasi_experimental
Evidence Strengthmedium — DML is a modern, credible approach that reduces bias from high-dimensional observed confounders, and the panel spans pre- and post-AI diffusion years; however, the small sample (45 listed firms), single-sector single-country focus, potential measurement error in AI and SCS constructs, and reliance on ignorability limit confidence in strong causal claims.
Methods Rigormedium — Methodologically sophisticated (DML, mediation decomposition, heterogeneity analysis) and leverages panel data over 12 years, but statistical power is limited by n=45, some identification assumptions are strong and not testable, and construct/measurement validity for AI intensity and supply-chain stability appears uncertain.
SampleFirm-level panel of 45 Chinese listed sports enterprises observed annually from 2012–2023, with measures of AI adoption, supplier-side and customer-side supply-chain stability, firm characteristics (including profitability) and other controls; listed firms only, sector-limited.
Themesadoption org_design human_ai_collab skills_training
IdentificationUses dual machine learning (DML) on a firm-year panel to estimate the causal effect of AI adoption on supply-chain stability, flexibly controlling for high-dimensional observed confounders and nuisance functions; causal interpretation relies on conditional ignorability (no unobserved time-varying confounders) and correct specification/estimation of nuisance models.
GeneralizabilitySmall sample size (45 firms) limits statistical power and external validity, Single sector (sports enterprises) — sector-specific dynamics may not generalize to manufacturing, services, or other industries, Listed firms only — excludes private firms and SMEs, which may adopt and experience AI differently, Single-country (China) — regulatory, market, and labor-institution context may drive results, Potential measurement/construct validity concerns for AI adoption and supply-chain stability metrics, Time period (2012–2023) may reflect particular stages of AI diffusion that differ elsewhere or later
Claims (6)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| AI significantly enhances supplier stability in sports enterprises (SE). Organizational Efficiency | positive | high | supplier stability (component of supply chain stability) |
n=45
statistically significant positive effect of AI on supplier stability
0.48
|
| AI notably reduces customer stability in sports enterprises (SE). Organizational Efficiency | negative | high | customer stability (component of supply chain stability) |
n=45
statistically significant negative effect of AI on customer stability
0.48
|
| Talent attraction is the primary mechanism through which AI affects supply chain stability in sports enterprises. Hiring | positive | medium | talent attraction as a mediator of AI's effect on supply chain stability |
n=45
talent attraction mediates AI's effect on supply chain stability (primary mechanism)
0.29
|
| Logistics efficiency does not mediate (fails to fulfill) the anticipated role in transmitting AI's effects to supply chain stability. Organizational Efficiency | null_result | medium | logistics efficiency as a mediator of AI's effect on supply chain stability |
n=45
no significant mediation via logistics efficiency
0.29
|
| The impact of AI on supply chain stability in sports enterprises exhibits heterogeneity by enterprise type and profitability status. Organizational Efficiency | mixed | medium | supply chain stability (SCS), analyzed across subgroups defined by enterprise type and profitability |
n=45
heterogeneous AI effects on supply chain stability by enterprise type and profitability
0.29
|
| In the digital economy, effective use of AI is crucial for maintaining supply chain stability in sports enterprises. Organizational Efficiency | positive | medium | overall supply chain stability (SCS) in sports enterprises |
n=45
effective AI use important for maintaining supply chain stability
0.29
|