How does artificial intelligence affect the carbon emission efficiency of resource-based cities? A nonlinear effect based on regional income inequality

The current development of resource-based cities is confronted with the dual predicament of sluggish economic growth and worsening environmental pollution. The wide application of artificial intelligence (AI) provides new insights and opportunities for their future paths. Based on the theoretical framework of "technology - income distribution - environment", this paper selects panel data of resource-based cities from 2006 to 2019 and nighttime light data with high spatial resolution, and uses various methods such as panel regression and threshold model to systematically investigate the intrinsic connection among AI, income inequality and carbon emission efficiency (CEE). The findings reveal that: (1) AI significantly enhances the CEE of resource-based cities, and this improvement has been achieved through advanced industrial structures and energy consumption intensities. (2) Income inequality is an important social channel that affects the carbon emission reduction effect of AI in resource-based cities, playing a nonlinear role of first promoting and then inhibiting. (3) This moderating effect shows significant heterogeneity due to the differences in the development stages of urban development. The research uncovers the underlying mechanism through which AI affects CEE from a social distribution perspective, and provides empirical evidence for formulating differentiated policies that balance carbon emission reduction and equity in resource-based cities at various stages of development.