A groundbreaking study conducted in China reveals that artificial intelligence (AI) development significantly enhances urban energy efficiency, offering a pivotal pathway toward achieving global sustainable development goals. The research, utilizing a fixed-effects model based on extensive prefecture-level data across Chinese cities, underscores AI's transformative potential in optimizing energy consumption and production patterns amidst evolving global economic and environmental landscapes.

The study identifies two primary mechanisms through which AI drives these efficiency gains: promoting green technological innovation and facilitating the rationalization of industrial structures. By leveraging advanced data analytics, predictive modeling, and automation, AI enables smarter decision-making in energy management, optimizes distribution networks, and accelerates the adoption of cleaner energy technologies. This systematic improvement in energy utilization is crucial for China, the world's largest energy consumer, and holds significant implications for other developing economies striving for sustainability.

Researchers found that AI's positive impact on energy efficiency is more pronounced in cities characterized by robust informal environmental regulations, suggesting that strong local governance and community oversight amplify the benefits of AI adoption. Conversely, the impact was less significant in areas with weaker environmental oversight. Furthermore, the study highlights that AI yields greater efficiency improvements in declining and regenerating resource-based cities compared to their growing and mature counterparts, indicating AI's potential as a catalyst for economic and environmental revitalization in transitioning regions.

China's AI industry, which exceeded 500 billion RMB by 2023 and encompasses over 4,500 enterprises, provides a unique large-scale context for this investigation. Unlike previous studies that often relied on proxies like industrial robots or patent counts, this research assessed the influence of AI enterprises directly, offering a more robust reflection of practical AI implementation. The use of a Data Envelopment Analysis (DEA) model, specifically the Charnes–Cooper–Rhodes (CCR) approach, allowed for a nuanced assessment of energy efficiency across multiple inputs and outputs, moving beyond simplistic energy-to-GDP ratios.

While acknowledging the energy-intensive nature of AI technologies—with data centers consuming substantial energy and the ICT sector's projected global energy use reaching 20% by 2030—the study emphasizes AI's net positive contribution to overall energy efficiency. The findings provide critical insights into the direct effects of AI on sustainable energy outcomes, urging policymakers to strengthen informal environmental governance and prioritize AI deployment in transitioning resource-based cities to fully realize these benefits and accelerate progress toward global sustainability.