Artificial Intelligence for Climate Change Mitigation: A Comprehensive Review of Decarbonisation Applications, Systemic Challenges, and Governance Imperatives

Artificial Intelligence (AI) models represent a critical frontier in accelerating global climate change mitigation efforts, offering capabilities that transcend traditional limits in optimization, prediction, and discovery. This review systematically synthesizes recent peer-reviewed literature detailing AI applications across four key mitigation domains: energy system optimization, industrial decarbonization (CCUS), regenerative agriculture, and low-carbon materials discovery. AI algorithms are shown to significantly enhance smart grid reliability and efficiency by predicting non-conventional energy output and maximizing energy storage utilization. Furthermore, in industrial settings, predictive analytics improves CCUS operational efficacy and subsurface modeling, despite facing significant infrastructure and data readiness barriers. Critically, the potential benefits of AI are counterbalanced by systemic risks. The AI Carbon Footprint Paradox, stemming from the high energy demands of training computation-intensive models, necessitates immediate trade-off analysis and corporate transparency.[1] Additionally, ethical concerns regarding data density bias and model opacity risk exacerbating global inequalities, as research disproportionately focuses on adaptation (64.4%) over mitigation (16.0%) and privileges the Global North.[2] Effective policy response and governance, particularly addressing the digital divide and fostering specialized expertise, are required to ensure AI deployment results in a net-positive impact on decarbonization targets.