Does Artificial Intelligence Contribute to Higher Per Capita Carbon Emissions and Ecological Footprint? Empirical Evidence from 61 Economies
Year: 2026
Type: preprint
Abstract: The impact of artificial intelligence on environmental quality has become a contentious issue in academic research. Clarifying the intrinsic relationship between the two is crucial for advancing global sustainable development. Employing a panel dataset of 61 countries worldwide from 1995 to 2023, this study systematically investigates the effects of artificial intelligence on per capita carbon emissions and ecological footprint. Empirical results demonstrate that artificial intelligence significantly increases both per capita carbon emissions and ecological footprint, thereby intensifying environmental pressure. Mechanism tests confirm that technological innovation and total-factor energy efficiency exert significant mediating effects in the process whereby AI influences the environment. Heterogeneity analysis indicates that the negative environmental effects of AI are primarily driven by general-purpose artificial intelligence, and such adverse impacts are more substantial in countries leading in AI development and upper-middle-income nations. Further research reveals that the environmental impact of AI exhibits nonlinearity subject to threshold constraints of technological innovation and total-factor energy efficiency. This study breaks through the conventional view that AI is inherently green, and uncovers the environmental costs behind its rapid expansion. The conclusions provide theoretical references for countries to formulate differentiated AI governance frameworks and green transition policies in line with their national conditions.
Source: Research Square
Authors: Qian Fei, JingFeng Zhao
Institution: Northwest University
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
SDG: Climate action
Open Access status: green
OpenAlex ID: W7236270665
