PaperlayerView exact revision on arXiv

Source-linked AI summary

Buyer Artificial Intelligence-Enabled Environmental Governance and Supplier Environmental Controversies: An Organizational Information Processing and Signaling

Yongchao Martin Ma, Xinya Guan

arXiv:2609.11391v1cs.CYcs.AIstat.AP

TL;DR

This study examines whether buyers’ AI-enabled environmental governance is associated with fewer environmental controversies among overseas suppliers, a spillover beyond prior focus on focal firms’ ESG outcomes. Using textual measures and panel data on 2,505 suppliers across 41 countries, it finds a negative next-year association that is stronger with higher supplier-country AI readiness and regulatory quality.

  • Problem

    Evidence has focused mainly on AI’s implications for focal firms’ ESG outcomes, leaving limited attention to whether AI-enabled governance influences environmental conduct beyond organizational boundaries.

    Source: §1 · p. 4, §2.1 · p. 8

  • Method

    The study measures buyer AI-enabled environmental governance through text analysis and analyzes supplier-year panel data with multidimensional fixed-effects models.

    Source: §1 · p. 5

  • Results

    Supplier exposure to buyer AI-enabled environmental governance is negatively associated with next-year environmental controversies, more strongly in countries with higher AI readiness and regulatory quality.

    Source: §1 · pp. 5–6

  • Takeaways & Limitations

    AI-enabled governance may complement and extend conventional monitoring and supplier oversight across global supply chains.

    Source: §1 · p. 6

    The stated relationships are associations concerning supplier environmental controversies, with the negative relationship conditioned by supplier-country AI readiness and regulatory quality.

    Source: §2.1 · p. 11, §2.4 · p. 13

Abstract

from arXiv · show

Environmental controversies in global supply chains pose significant risks for global buyers. This study examines whether overseas suppliers' exposure to buyers' artificial intelligence (AI)-enabled environmental governance reduces supplier environmental controversies. Drawing on organizational information processing theory and signaling theory, we investigate how suppliers' exposure to AI-enabled governance influences their environmental controversies and the institutional contingencies under which this effect varies. Using text analysis to measure buyer AI-enabled environmental governance, we analyze panel data on 2,505 suppliers of U.S.-listed firms across 41 countries from 2020 to 2024 with multidimensional fixed-effects models. We find that suppliers' exposure to buyer AI-enabled environmental governance is negatively associated with supplier environmental controversies in the following year. This negative relationship is stronger in supplier countries with higher AI readiness and regulatory quality. The study contributes to research on AI-enabled sustainability governance and sustainable supply chain risk management.

1. Introduction

The study asks whether buyers’ AI-enabled environmental governance spills over to foreign suppliers, addressing limited evidence beyond focal firms’ ESG outcomes. It finds a negative association with suppliers’ next-year environmental controversies, stronger where supplier-country AI readiness and regulatory quality are higher.

1. Introduction

  • Traditional governance struggles to provide timely, reliable oversight across supply chains with fragmented information, limited visibility, and heterogeneous institutions.AI offers scalable ways to gather, interpret, and synthesize complex information across organizational and national boundaries.Source: §1 · p. 3
  • Supplier exposure to buyer AI-enabled environmental governance is negatively associated with next-year environmental controversies, with stronger associations in countries with higher AI readiness and regulatory quality.The study uses sentence-level analysis of buyers’ 10-K filings and a panel of 2,505 suppliers across 41 countries.Source: §1 · p. 5
  • The study extends AI-sustainability research beyond focal firms’ ESG outcomes by examining cross-border governance spillovers to foreign suppliers.It also identifies supplier-country AI readiness and regulatory quality as contingencies.Source: §1 · pp. 4, 6
  • The authors recommend that buyers adopt and disclose AI-enabled environmental governance and that supplier-country policymakers invest in AI infrastructure and regulatory quality.They also highlight cross-border collaboration between firms and policymakers in improving oversight.Source: §1 · pp. 6–7

2. Literature Review and Hypothesis Development

The section develops a framework in which buyer AI-enabled governance may reduce supplier environmental controversies through information processing and oversight signals. It proposes that this negative association is stronger in supplier countries with higher AI readiness and regulatory quality.

2.1 Sustainable Supply Chain Risks and AI-Enabled Sustainability Governance

  • AI-enabled governance may improve information processing across dispersed supply networks and signal stronger oversight, making supplier controversies less likely.The argument addresses fragmented environmental information that traditional supply-chain practices struggle to gather and process.Source: §2.1 · pp. 7, 10, 11

2.3 The Moderating Role of Supplier-Country AI Readiness

  • Higher supplier-country AI readiness is expected to strengthen the negative relationship by supporting data infrastructure, real-time communication, integration, traceability, and coordination.Lower readiness may constrain buyers’ information-processing effectiveness through weaker infrastructure and technological compatibility.Source: §2.3 · pp. 11–12

2.4 The Moderating Role of Supplier-Country Regulatory Quality

  • Higher supplier-country regulatory quality is expected to strengthen the negative relationship by reinforcing oversight signals, compliance expectations, and environmental information transparency.Weaker regulatory support may reduce signal credibility and constrain information availability.Source: §2.4 · pp. 12–13

3. Methodology

The study assembles a supplier-year panel of foreign suppliers to non-financial, U.S.-listed firms and measures supplier controversies and buyers’ AI-enabled environmental governance using text and incident data. The design aggregates buyers’ governance exposure to suppliers and uses country-level AI readiness and regulatory quality as moderators.

3.1 Sample and Data Sources

  • The unbalanced panel contains 5,620 supplier-year observations from 2,505 foreign suppliers in 41 countries between 2020 and 2024.The sample focuses on foreign suppliers of non-financial, U.S.-listed firms; buyers’ governance exposure is aggregated at the supplier-year level using buyer-importance weights.Source: §3.1 · pp. 13–15

3.2.1 Dependent Variable

  • Supplier environmental controversies are measured as each supplier’s annual count of media-reported negative environmental incidents tracked by RepRisk.RepRisk uses broad multilingual news coverage and analyst verification; robustness checks also use an incident sum weighted by severity and reach.Source: §3.2.1 · p. 18

3.2.2 Key Independent Variable

  • Buyer AI-enabled environmental governance is measured through textual analysis of 10-K disclosures about AI applied to environmental governance activities.The measure targets co-occurring AI, environmental, and governance terms within sentences rather than standalone keyword frequencies.Source: §3.2.2 · pp. 18–19
  • The text measure uses AI, environmental, and governance lexicons, with AI terms supplemented for generative AI and environmental and governance terms manually validated in the 10-K corpus.Ambiguous AI acronyms were excluded to reduce false positives.Source: §3.2.2 · pp. 19–20

Appendix A.

The section describes how buyer AI-enabled environmental governance is measured and aggregated, defines country-level moderators and controls, and specifies fixed-effects models for supplier environmental incidents.

Appendix A.

  • Buyer governance is measured as qualifying AI–environment–governance sentences per 1,000 report sentences, then supplier-aggregated using relationship-duration weights.Triple co-occurrence reduces false positives from standalone topic discussions; a random subset was manually validated, and buyer net assets were considered as an alternative weighting scheme.Source: Appendix A · pp. 20–21

3.2.3 Moderating Variables

  • AI readiness uses the annual Oxford Insights Government AIR index, while regulatory quality uses the Worldwide Governance Indicators’ Regulatory Quality measure.The AIR index combines government, technology-sector, and data-and-infrastructure pillars; higher WGI values indicate stronger regulatory quality.Source: §3.2.3 · pp. 21–22

3.2.4 Control Variables

  • Controls cover supplier resources, non-U.S. customer exposure, buyer characteristics and monitoring, and supplier-country institutions and GDP.Supplier controls include size, age, net assets, and leverage; buyer controls include environmental performance, controversies, monitoring, and net assets.Source: §3.2.4 · pp. 22–23

3.3 Model Specification

  • PPML models next-year supplier incident coverage against lagged buyer AI-governance exposure and controls, with supplier-industry, supplier-country, and year fixed effects.The dependent variable is a nonnegative, right-skewed count with many zeros; the model uses one-year-lagged predictors.Source: §3.3 · pp. 23–24
  • Moderated specifications include exposure, each country moderator, and their mean-centered interaction simultaneously to distinguish moderation from direct moderator effects.The moderators are supplier-country AI readiness and regulatory quality.Source: §3.3 · p. 24

4. Results

Exposure to buyer AI-enabled environmental governance is negatively associated with suppliers’ next-year environmental controversies, and the relationship is stronger in countries with higher AI readiness and regulatory quality. Supplementary analyses report consistent results across alternative measures, samples, and estimation approaches, with an instrumental-variable analysis providing additional support.

4.1 Descriptive Statistics

  • Supplier controversies range from 0 to 202 and exposure to buyer AI-enabled environmental governance from 0 to 4.66 in the sample.The exposure–controversies pairwise correlation is r = −0.042 (p < 0.01).Source: §4.1 · pp. 24–25

4.2 Main Results

  • β = −0.282 (p < 0.001): exposure to buyer AI-enabled environmental governance is negatively associated with suppliers’ next-year environmental controversies.The estimate is from Model 2, which introduces the primary independent variable.Source: Table 4 · p. 27, §4.2 · p. 29
  • The negative association is stronger in countries with higher AI readiness and regulatory quality.Figures 2 and 3 illustrate the respective moderating relationships.Source: Figure 2 · p. 28, Figure 3 · p. 28, §4.2 · p. 29

4.3 Robustness Checks

  • Supplementary analyses using alternative measures, samples, and specifications consistently corroborated the primary results.Detailed outcomes are reported in Online Appendix B.Source: §4.3 · p. 30

4.3.1 Alternative Measures for Key Variables

  • Severity/reach-weighted controversies and asset-weighted exposure retained baseline results; full-text exposure was directionally consistent, but its main effect fell to marginal significance.The weighted controversy measure reflects incident severity and information-source reach.Source: §4.3.1 · p. 30

4.3.2 Alternative Samples

  • Results remained largely consistent across alternative samples; among 662 supplier-years with controversies, the main effect remained negative and significant, while interaction effects attenuated.In that restricted sample, the AI-readiness interaction became insignificant and the regulatory-quality interaction remained significant only at the 10% level.Source: §4.3.2 · p. 31

4.3.3 Alternative Estimation Methods

  • Negative-binomial estimates were qualitatively consistent with baseline PPML results, suggesting findings were not driven by the estimator’s distributional assumptions.The negative-binomial models included supplier-industry and year fixed effects.Source: §4.3.3 · p. 31

4.3.4 Endogeneity

  • Using state AI bills and AI Geographic Exposure as instruments, second-stage exposure remained negatively associated with supplier controversies (β = −2.026, p < 0.05).The reported diagnostics reject underidentification and suggest weak-instrument concerns are unlikely.Source: §4.3.4 · p. 32

5. Discussion

Buyer AI-enabled environmental governance is associated with fewer supplier environmental controversies, with stronger associations in countries with greater AI readiness and regulatory quality. The discussion develops the information-processing and signaling mechanisms, practical implications, and limitations of this cross-border governance relationship.

5. Discussion

  • Supplier exposure to buyer AI-enabled environmental governance is negatively associated with environmental controversies, and the association is stronger where supplier-country AI readiness and regulatory quality are higher.The authors interpret AI readiness as supporting information processing and regulatory quality as strengthening governance signals and oversight.Source: §5 · pp. 32–33

5.1 Theoretical Contributions

  • Buyer AI-enabled governance may reduce supplier controversies by improving environmental information processing across dispersed networks and signaling heightened scrutiny; supplier-country AI readiness and regulatory quality strengthen these mechanisms.AI readiness provides technical conditions for information processing, while regulatory quality supports signal credibility and enforceability.Source: §5.1 · pp. 33–34
  • The study measures buyer AI-enabled environmental governance by identifying joint AI, environmental-issue, and environmental-governance disclosures in SEC 10-K filings, then weighting the measure across buyer–supplier relationships.The measure is based on sentence-level textual analysis of Items 1 and 7.Source: §5.1 · p. 34
  • AI-enabled governance is framed as a scalable complement to conventional supplier oversight that supports continuous information processing, broader network visibility, and earlier, more systematic risk identification.Conventional audits, monitoring, termination, and training can be resource-intensive, episodic, and difficult to scale across dispersed networks.Source: §5.1 · pp. 34–35

5.2 Managerial Implications

  • Buyer firms are advised to disclose AI-enabled environmental governance and use it alongside audits, training, and corrective-action systems rather than as a replacement.The authors suggest using AI tools to monitor environmental information continuously and direct oversight toward high-risk suppliers and regions.Source: §5.2 · p. 35
  • Supplier-country policymakers are urged to strengthen AI readiness, regulatory quality, and cross-border coordination to support more transparent, reliable, and actionable environmental oversight.The discussion highlights digital and data infrastructure, disclosure standards, enforcement, monitoring institutions, shared data standards, and interoperable systems.Source: §5.2 · pp. 35–36

5.3 Limitations and Future Research

  • The findings are limited to suppliers of U.S. buyers and are primarily correlational, with some specifications unable to include firm fixed effects and therefore potentially subject to unobserved firm characteristics.The authors suggest examining other buyer-country contexts and using stronger causal designs; surveys or interviews could directly investigate proposed mediating mechanisms.Source: §5.3 · pp. 36–37
  • Without transaction-level or relationship-specific data, the study cannot distinguish whether governance is more effective for suppliers with greater transaction volume or dependence.Future research could combine buyer–supplier links with contract or procurement data.Source: §5.3 · p. 37

6. Conclusion

The study finds a negative association between buyer AI-enabled environmental governance and subsequent supplier environmental controversies among overseas suppliers of U.S.-listed firms. This association is stronger in supplier countries with higher AI readiness and regulatory quality.

6. Conclusion

  • Buyer AI-enabled environmental governance is negatively associated with subsequent supplier environmental controversies in the 2020–2024 panel of overseas suppliers of U.S.-listed firms.The analysis uses multidimensional fixed-effects models.Source: §6 · p. 37
  • The stronger association in countries with higher AI readiness and regulatory quality is consistent with supportive digital infrastructure and more credible, effective environmental oversight.The paper concludes that buyer-side AI-enabled governance can generate cross-border governance spillovers and may contribute to more sustainable supply chain management.Source: §6 · p. 38
Loading 2609.11391v1…