Digitalization, Energy Use, and Carbon Emission Efficiency: Testing the Digital Paradox

The rapid expansion of digital technologies is reshaping economic and energy systems, but its implications for carbon efficiency remain uncertain. This study examines whether digitalization improves carbon emission efficiency (CEE), investigates the mediating role of green innovation, and tests whether ICTI conditions the digitalization–CEE relationship. Using a panel of 53 economies with varying energy intensities over 2002–2022, CEE is estimated through the Super-Efficiency Slack-Based Measure (SE-SBM), which incorporates energy consumption and carbon dioxide emissions as environmental inputs and undesirable outputs, respectively. Panel-corrected standard errors, Pooled Mean Group, System-GMM, and Driscoll–Kraay estimators are employed for robust inference. The results show that digitalization significantly reduces CEE (β = −0.1677), indicating that the energy and carbon pressures associated with expanding DII outweigh initial efficiency gains. In particular, the findings suggest that rising electricity and energy demand associated with data centers, communication networks, and ICT-intensive activities can intensify environmental pressure. Green innovation positively improves CEE (β = 0.0089) and partially mediates the adverse effect of digitalization, although digitalization itself significantly reduces green innovation (β = −0.0820). Moreover, the digitalization-ICTI interaction is negative and significant (β = −0.6864), indicating that higher ICTI amplifies the adverse environmental effect of digital expansion. Institutional quality improves CEE, whereas a resource-intensive industrial structure reduces it. A U-shaped relationship between economic growth and CEE is also confirmed. The findings highlight the need for energy-efficient DII, efficient ICT utilization, green innovation, stronger institutions, and sustainable industrial transformation to reconcile digital development with environmental sustainability.