← Papers

Chinese manufacturers that invest in AI show materially lower carbon intensity, largely by reshaping governance, decarbonizing production and redirecting investment to green assets; the emission gains are concentrated in firms with dispersed supply chains, in environmentally sensitive industries and in regions with underdeveloped factor markets.

Artificial Intelligence Innovation, Internal Structure Optimization and Corporate Carbon Emission Reduction: Experience from China

Xingxing Lu, Lianying Liao, Xiaojuan Luo, Bing Zhao · April 02, 2026 · Sustainability

openalex correlational medium evidence 7/10 relevance DOI Source PDF

Using a large panel of Chinese listed manufacturers, the study finds that firms' AI innovation activity is associated with significantly lower carbon emission intensity, working through improved low-carbon governance, cleaner production process upgrades, and reallocation of investment to green assets, with stronger effects where executives are environmentally minded and government environmental attention is high.

Against the dual backdrop of the global drive toward carbon peaking and carbon neutrality, a core pillar of the United Nations Sustainable Development Goals (SDGs), and the accelerated integration of new-generation digital technologies into sustainable production practices, this study employs a micro-level perspective to systematically explore how AI innovation optimizes organizational, production, and investment structures to enable corporate low-carbon development. The study sample comprises 21,428 firm-year observations from Chinese A-share listed manufacturing companies over the period of 2010–2022. The results show that AI innovation can significantly reduce corporate carbon emission intensity, specifically achieving corporate low-carbon development through three paths: optimizing low-carbon organizational governance, upgrading emission-reducing production processes, and reorienting investment toward green assets. Further analysis shows that executives’ green cognition and government environmental attention together constitute dual internal and external driving forces for corporate carbon emission reduction. Heterogeneity analysis reveals that the emission reduction effect of AI innovation is more significant for enterprises with a low supply chain concentration, those in high-environmental-sensitivity industries, and those located in regions with underdeveloped factor markets. From the micro-perspective of corporate sustainable low-carbon development, this study offers further theoretical support and empirical evidence for regulators aiming to optimize AI innovation incentives, improve sustainable environmental governance, and advance global sustainable industrial development.

Summary

Main Finding

AI innovation at the firm level significantly reduces corporate carbon-emission intensity. The paper shows that firms leverage AI to achieve low‑carbon development via three micro-level channels—optimizing organizational governance for low carbon, upgrading production processes to cut emissions, and reallocating investment toward green assets. Executives’ green cognition and government environmental attention jointly drive these effects, and the emission-reduction impact of AI is stronger in firms with low supply‑chain concentration, in environmentally sensitive industries, and in regions with less developed factor markets.

Key Points

Magnitude and direction
- AI innovation → statistically and economically significant decline in firm-level carbon-emission intensity.
Mechanisms (three mediation paths)
- Organizational governance: AI helps redesign decision-making, monitoring, and incentives to prioritize low-carbon outcomes.
- Production processes: AI enables process optimization and emission-reducing technological upgrades.
- Investment reorientation: AI supports shifting capital toward green assets and projects.
Dual drivers
- Internal: executives’ green cognition strengthens the translation of AI innovation into emission reductions.
- External: government environmental attention amplifies firm responses.
Heterogeneity
- Stronger AI-induced emission reductions for firms with low supply‑chain concentration.
- Larger effects in high-environmental-sensitivity industries.
- Greater impacts in regions with underdeveloped factor markets.
Policy relevance
- Findings support targeted AI-innovation incentives and environment-focused governance to accelerate corporate decarbonization.

Data & Methods

Sample
- 21,428 firm‑year observations from Chinese A‑share listed manufacturing firms, 2010–2022.
Outcome and main predictor
- Dependent variable: firm carbon-emission intensity (firm-level emissions scaled by output or revenue—reported by authors).
- Key independent variable: firm-level AI innovation (measured by the authors using firm AI-innovation indicators).
Empirical approach (reported analyses)
- Micro‑level panel analysis assessing the relationship between AI innovation and carbon intensity.
- Mediation tests to identify the three organizational/production/investment channels.
- Moderation analyses to test the roles of executives’ green cognition and government environmental attention.
- Heterogeneity analyses across supply‑chain concentration, industry environmental sensitivity, and regional factor-market development.
- Robustness checks and additional analyses (as reported) to support the findings.

Implications for AI Economics

AI as a decarbonizing general-purpose technology
- Empirical evidence that AI adoption can materially lower firms’ carbon intensity via organizational, process, and capital-allocation changes.
Policy design
- Complement carbon and innovation policy: subsidies or tax incentives for AI R&D targeted at emissions-reducing applications; integrate AI strategy into environmental regulation and industrial policy.
- Strengthen external drivers: clearer environmental signals and enforcement amplify firms’ use of AI for decarbonization.
Firm strategy
- Managers should align AI investments with governance reforms and green-capital allocation to realize emissions benefits; managerial green cognition matters for realizing returns from AI.
Market and institutional context matters
- AI’s effectiveness for emissions reduction depends on supply‑chain structure, industry exposure to environmental regulation, and local factor-market development—policies should be tailored accordingly.
Research directions
- Causal identification of AI’s carbon effects (e.g., exogenous shocks or instruments).
- Cross-country comparisons and non‑listed firms.
- Interaction of AI adoption with energy-sector decarbonization, rebound effects (AI energy use), and long‑run productivity-emissions tradeoffs.

Assessment

Paper Typecorrelational Evidence Strengthmedium — Large longitudinal sample, fixed effects and robustness/mediation tests increase credibility, but no clearly stated exogenous source of variation (instrument, discontinuity, or randomized shock) is described to rule out endogeneity (reverse causality and omitted confounders) or measurement error in AI and emissions variables, limiting causal certainty. Methods Rigormedium — The study appears to apply standard and appropriate econometric tools for observational panel data (fixed effects, mechanism tests, interactions, heterogeneity), with a large sample and multi-year coverage; however, the absence of a strong identification strategy (e.g., IV, natural experiment) and possible concerns about variable measurement reduce methodological rigor relative to best-practice causal work. SampleFirm-level panel of Chinese A-share listed manufacturing firms, 2010–2022, covering 21,428 firm-year observations; key variables likely include firm-level measures of AI innovation (e.g., AI-related patents or R&D), corporate carbon emission intensity (firm-level emissions per output), executive green cognition and government environmental attention proxied by text or policy indicators, and standard firm controls. Themesinnovation org_design governance IdentificationPanel regression analysis on 21,428 firm-year observations of Chinese A-share listed manufacturing firms (2010–2022) using firm and year fixed effects, control variables, robustness checks, mediation tests to probe three hypothesized causal paths (organizational governance, production process upgrading, green investment reorientation), interaction analyses with proxies for executives' green cognition and government environmental attention, and heterogeneity analysis across supply-chain concentration, industry environmental sensitivity, and regional factor-market development. GeneralizabilityLimited to publicly listed manufacturing firms in China — may not generalize to private firms, services, or firms in other countries, Chinese regulatory and policy context (e.g., government environmental attention) may drive effects that differ elsewhere, AI innovation and carbon emissions measurement choices (patents, text proxies, estimated emissions) may not capture all relevant activity or apply in other datasets, Findings pertain to 2010–2022 and may not hold as AI technology and regulatory regimes evolve, Potential bias toward larger, more formal firms that are listed and report required data

Claims (9)

Claim	Direction	Confidence	Outcome	Details
AI innovation can significantly reduce corporate carbon emission intensity. Organizational Efficiency	negative	high	corporate carbon emission intensity	n=21428 0.5
AI innovation achieves corporate low-carbon development by optimizing low-carbon organizational governance. Organizational Efficiency	positive	high	corporate carbon emission intensity (mediated via organizational governance changes)	n=21428 0.3
AI innovation achieves corporate low-carbon development by upgrading emission-reducing production processes. Organizational Efficiency	positive	high	corporate carbon emission intensity (mediated via production process upgrades)	n=21428 0.3
AI innovation achieves corporate low-carbon development by reorienting investment toward green assets. Organizational Efficiency	positive	high	corporate carbon emission intensity (mediated via investment reorientation toward green assets)	n=21428 0.3
Executives’ green cognition and government environmental attention together constitute dual internal and external driving forces for corporate carbon emission reduction. Organizational Efficiency	negative	high	corporate carbon emission intensity / carbon emission reduction	n=21428 0.3
The emission-reduction effect of AI innovation is more significant for enterprises with a low supply chain concentration. Organizational Efficiency	negative	high	corporate carbon emission intensity (differential effect by supply chain concentration)	n=21428 0.3
The emission-reduction effect of AI innovation is more significant for firms in high-environmental-sensitivity industries. Organizational Efficiency	negative	high	corporate carbon emission intensity (differential effect by industry environmental sensitivity)	n=21428 0.3
The emission-reduction effect of AI innovation is more significant for firms located in regions with underdeveloped factor markets. Organizational Efficiency	negative	high	corporate carbon emission intensity (differential effect by regional factor market development)	n=21428 0.3
The study sample comprises 21,428 firm-year observations from Chinese A-share listed manufacturing companies over 2010–2022. Other	null_result	high	sample composition (firm-year observations)	n=21428 0.5