The Commonplace
Home Papers Evidence Explore Trends Syntheses Digests References Docs 🎲 Workforce Futures
← Papers
Direction, evidence grade, and study type are AI-generated labels (gpt-5-mini), not human-verified. Syntheses are LLM-written. "Tensions" are machine-detected candidates, not confirmed contradictions. A research-acceleration tool, not peer review. How this is built →

Firms' surface-level focus on AI can crowd out environmental risk management, but deeper AI capabilities and AI-driven innovation reinforce green practices; financing frictions and uncertainty shape these trade-offs, while subsidies and tax incentives help align AI investment with greener outcomes.

The Economic Mechanisms of Artificial Intelligence Resources Affecting Green Risk Management: Empirical Evidence from Chinese Listed Firms
Yi Huang, Zhe Zhang, Sui Sun · February 24, 2026 · International Journal of Economic Sciences
openalex correlational medium evidence 7/10 relevance Full text usable extracted full text DOI Source PDF

Structured author observations

Linked only from stored provider relations; the raw author line above is never matched by name.

OpenAlex

Latest observation:

  1. Yi Huang provider ID
  2. Zhe Zhang provider ID
  3. Sui Sun provider ID

Semantic Scholar

Latest observation:

  1. Yi Huang provider ID
  2. Zhe Zhang provider ID
  3. Sui Sun provider ID
Using firm-level observational measures, the paper finds that firms' AI attention is associated with weaker green risk management, while AI depth and AI-driven innovation strengthen GRM, with financing, environmental uncertainty, and policy incentives moderating these relationships.

Citation observations

Cumulative provider counts captured on specific dates; providers are never combined.

Against the backdrop of the digital economy, artificial intelligence (AI) has become a critical strategic resource for firms. Beyond enhancing economic performance, AI has opened new pathways for addressing environmental externalities. Drawing on the resource-based view (RBV), this paper examines the economic mechanisms through which AI resources influence firms’ green risk management (GRM) practices. The empirical results show that AI attention has a negative impact on GRM, whereas AI depth and AI-driven innovation exert positive effects. Further analysis indicates that AI technology attention and software depth are the main factors driving the negative impact. A resource crowding-out effect exists between AI attention and green attention, but this effect is weakened when financing constraints are low. Although no crowding-out effect is observed between AI depth and green depth, environmental uncertainty promotes the emergence of such an effect. At the same time, green innovation is identified as an indispensable mediating variable in the process through which AI attention and AI depth influence GRM. Further analysis confirms that environmental subsidies and tax incentives, as key economic policy instruments, can complement AI resources and effectively promote green risk management.

Summary

Main Finding

Heterogeneous AI resources affect firms’ green risk management (GRM) differently. AI attention (management/strategic focus on AI) reduces GRM, while AI depth (capitalized AI assets per employee) and AI-driven innovation improve GRM. The negative effect of AI attention operates through a resource crowding-out of green attention (mitigated when financing constraints are low), whereas AI depth generally complements green capabilities except under high environmental uncertainty. Green innovation mediates part of the positive effects of AI depth/attention on GRM. Environmental subsidies and tax incentives strengthen AI’s positive role in GRM.

Key Points

  • Three AI resource dimensions (resource-based view):
    • AI attention (Ai_A): measured by frequency of AI-related keywords in disclosures — captures cognitive/attention resources.
    • AI depth (Ai_D): AI-related capital intensity per employee — captures foundational, integrated assets.
    • AI innovation (Ai_I): presence of AI-related patents (binary) — captures cumulative/innovative resources.
  • Main empirical results:
    • Ai_A has a statistically significant negative effect on firms’ GRM.
    • Ai_D and Ai_I have statistically significant positive effects on GRM.
  • Mechanisms and heterogeneity:
    • AI attention → resource crowding-out of green attention (reduces GRM); effect weaker when financing constraints are low.
    • No baseline crowding-out between AI depth and green depth, but environmental uncertainty can induce crowding-out.
    • Green innovation is an important mediator linking AI attention/depth to GRM.
    • Decomposition shows AI technology attention and software depth mainly drive the negative effects associated with attention.
  • Policy interaction:
    • Environmental subsidies and tax incentives complement AI resources and enhance firms’ GRM.

Data & Methods

  • Sample: Chinese A-share listed firms (Shanghai and Shenzhen exchanges), 2010–2023; final panel of 29,285 firm-year observations after exclusions.
  • Data sources: corporate annual reports, CSR reports, patent data from CNRDS; financial/governance variables from CSMAR; regional/macro variables from China Statistical Yearbook.
  • Dependent variable: GRM — composite index based on ISO 31000-inspired dimensions (risk identification, assessment, management measures, prevention, and effectiveness) composed of 32 sub-indices. Weights computed via Entropy–CRITIC; indices normalized before aggregation.
  • Key independent variables:
    • Ai_A = ln(1 + count of AI-related keywords in disclosures).
    • Ai_D = book value of AI-related assets / number of employees.
    • Ai_I = indicator = 1 if firm holds AI-related patent(s), else 0.
  • Controls: firm age, size, ROE, leverage, revenue growth, cash ratio, board size, investment intensity, etc.
  • Estimation approach: panel econometric models (benchmark regressions with controls and robustness checks). Tests include:
    • Mechanism tests (mediation analysis for green innovation).
    • Interaction/moderation tests (financing constraints, environmental uncertainty).
    • Decomposition/subgroup analyses (e.g., technology attention, software depth).
  • Data cleaning: excluded ST/*ST firms, removed observations with missing key variables, applied 1% Winsorization.

Implications for AI Economics

  • Resource heterogeneity matters: Policymakers and managers should distinguish between attention-level AI investments (which can crowd out environmental focus) and deep, asset-based AI investments that are more likely to generate sustainable green outcomes.
  • Managerial strategy: Firms should avoid attention-only AI strategies (publicity, superficial focus); prioritize embedding AI into operations and R&D to realize GRM benefits. Strengthen governance to prevent managerial myopia toward AI at the expense of green priorities.
  • Financing and environment conditions: Low financing constraints reduce the crowding-out risk of AI attention; environmental uncertainty can negate the complementarity of AI depth with green capabilities. Financial policies that ease financing for green-directed AI investments can be effective.
  • Policy design: Environmental subsidies and tax incentives can synergize with AI investments to improve GRM. Governments should target incentives to encourage deep integration of AI into green innovation and operations (not merely attention/PR).
  • Research and measurement: The mediating role of green innovation underscores the value of supporting AI-enabled green R&D (e.g., AI for emissions monitoring, predictive maintenance, circular systems).

Limitations noted by the study (and avenues for future work): single-country (China) listed-firm sample limits generalizability; some AI measures (keyword-based attention, binary patents) are coarse proxies; causal identification could be strengthened in future research; further work could examine energy/environmental costs of AI deployment and cross-country policy interactions.

Assessment

Paper Typecorrelational Evidence Strengthmedium — The paper documents consistent associations and conducts mediation and heterogeneity checks, but relies on observational variation and proxy-based measures of AI resources, leaving key endogeneity concerns (reverse causality, omitted variables, measurement error) insufficiently addressed for strong causal claims. Methods Rigormedium — The authors appear to construct nuanced firm-level measures (text-based attention, depth metrics), run multivariate regressions, and explore mechanisms and moderators (mediation, interactions, policy variables), which is rigorous for correlational work; however, the lack of exogenous variation, limited discussion of identification tests (e.g., instruments, difference-in-differences, or IV), and reliance on proxies reduce overall methodological rigor. SampleFirm-level observational data (likely publicly listed firms) combining corporate disclosures/texts, financial statements, patent or software development measures, and policy/subsidy data to construct AI attention, AI depth, AI-driven innovation, and green risk management (GRM) indicators; panel or cross-sectional structure implied but exact country, period and sample size not specified in the excerpt. Themesorg_design innovation IdentificationObservational firm-level regression analysis using variation in constructed AI measures (AI attention from text mining, AI depth from software/patent/deployment measures, AI-driven innovation), with controls, mediation tests, and interaction (heterogeneity) analyses; no clear exogenous instrument or natural experiment is described, so identification is associational rather than causal. GeneralizabilityLikely country- and sample-specific (e.g., listed firms), so results may not generalize to small/private firms or other institutional contexts, Measures rely on textual proxies and patent/software indicators which may imperfectly capture actual AI capability or deployment, Sectoral composition may bias results if certain industries dominate AI or green practices, Time-period specific to when AI adoption and environmental policies were evolving, limiting transportability to other periods, Associational design limits causal generalizability across contexts with different confounders or policy regimes

Claims (10)

ClaimDirectionOutcomeConfidence & EvidenceDetails
AI attention has a negative impact on firms' green risk management (GRM). Organizational Efficiency negative green risk management (GRM)
Reading fidelity high
Study strength medium
not reported
0.3
AI depth has a positive effect on firms' green risk management (GRM). Organizational Efficiency positive green risk management (GRM)
Reading fidelity high
Study strength medium
not reported
0.3
AI-driven innovation exerts a positive effect on firms' green risk management (GRM). Organizational Efficiency positive green risk management (GRM)
Reading fidelity high
Study strength medium
not reported
0.3
AI technology attention and software depth are the main factors driving the observed negative impact of AI attention on GRM. Organizational Efficiency negative green risk management (GRM)
Reading fidelity high
Study strength medium
not reported
0.3
There is a resource crowding-out effect between AI attention and green attention (AI attention crowds out green attention). Task Allocation negative green attention / allocation of resources to green activities
Reading fidelity high
Study strength medium
not reported
0.3
The crowding-out effect between AI attention and green attention is weakened when firms face low financing constraints. Task Allocation positive strength of crowding-out between AI attention and green attention
Reading fidelity high
Study strength medium
not reported
0.3
No crowding-out effect is observed between AI depth and green depth overall, but environmental uncertainty promotes the emergence of such an effect. Task Allocation mixed crowding-out between AI depth and green depth (resource allocation)
Reading fidelity high
Study strength medium
not reported
0.3
Green innovation is an indispensable mediating variable through which AI attention and AI depth influence firms' green risk management. Organizational Efficiency mixed green risk management (GRM) with green innovation as mediator
Reading fidelity high
Study strength medium
not reported
0.3
Environmental subsidies and tax incentives can complement AI resources and effectively promote green risk management. Organizational Efficiency positive green risk management (GRM)
Reading fidelity high
Study strength medium
not reported
0.3
AI has become a critical strategic resource for firms and has opened new pathways for addressing environmental externalities. Innovation Output positive role of AI as strategic resource and its potential for environmental externality mitigation
Reading fidelity high
Study strength speculative
not reported
0.05

Notes