Combating ESG Greenwashing Through AI Models: Evidence From Disaggregated AI Technologies, Mechanisms, and Thresholds

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:The disaggregated AI technology analysis and threshold effects offer a novel theoretical lens for studying AI-governance interactions.

🏢実務担当者:Firms can leverage AI models to monitor and improve ESG reporting transparency, particularly in high-pollution or technology-intensive sectors.

🏛政策担当者:Regulators should consider integrating AI analytical capabilities into greenwashing detection frameworks and designing differentiated support for AI adoption in ESG.

This study examines how artificial intelligence language models influence corporate environmental, social, and governance greenwashing () behavior, utilizing panel data from Chinese listed firms spanning 2012–2022. We investigate the mechanisms, technological disaggregation, threshold dynamics, and contextual heterogeneity characterizing this relationship. Employing a large language model to construct refined measures of AI adoption, our empirical analysis yields four principal findings. First, AI language models reduce . Second, technological disaggregation demonstrates substantial heterogeneity across AI subfields: machine learning and planning‐decision systems exert the strongest constraining effects on , while other AI categories show limited influence. Third, mechanism analysis reveals that AI constrains through two primary transmission channels, workforce skill restructuring and firm performance enhancement. Fourth, threshold regression analysis reveals critical nonlinearities: the AI– GW ESG relationship exhibits structural shifts contingent upon environmental regulation intensity and green technology sophistication. Fifth, subsample analysis stratified by pollution intensity, ownership structure, and technological orientation shows that AI's deterrent effect intensifies among heavily polluting enterprises, non‐state‐owned firms, and technology‐intensive sectors—contexts where disclosure credibility faces maximum scrutiny and organizational absorptive capacity enables effective AI integration. These findings suggest that recognizing AI's potential for environmental governance requires coordinated policy interventions: promoting AI deployment in environmentally critical domains, strengthening regulatory frameworks that leverage AI's analytical capabilities, investing in complementary human capital and digital infrastructure, and designing differentiated support mechanisms calibrated to firm heterogeneity. Firms can leverage on Ai models to boost transparency and Accountability of ESG reporting. Regulators and policymakers can investigate developing guidelines on the usage of AI technology to curb ESG greenwashing and improve sustainable practices.

• semanticscholar https://doi.org/10.1002/csr.70689first seen 2026-05-26 05:00:33 · last seen 2026-05-27 05:01:15