The energy system, climate change, air pollution, and their associated effects on human health are intricately interconnected. This research introduces an innovative AI-powered integrated model that combines optimization techniques for energy systems with machine learning approaches, including XGBoost for PM 2 . 5 prediction and Prophet models for forecasting weather data and energy demand. This hybrid approach enhances predictive accuracy and enables a comprehensive evaluation of the co-benefits of mitigating carcinogenic impact, climate change, and economic outcomes in urban energy transitions. Findings suggest that, although effective economic management can reduce PM 2 . 5 concentrations to comply with national and EPA regulations, achieving WHO’s more stringent guideline of 5 μg/m 3 by 2045 necessitates a net-zero emissions target. The ambitious "Cheetah" scenario attains the most significant decrease in PM 2 . 5 , nearly eradicating seasonal variations and potentially decreasing cancer-related fatalities by 92%. Conversely, less ambitious scenarios such as “Caracal” and “Hyena” forecast only incremental advancements, failing to meet WHO criteria by 2050 and indicating ongoing health hazards. Moreover, AI-driven forecast analyses, particularly using Prophet models, underscore the imperative for robust actions to stabilize PM 2 . 5 throughout seasonal and high-emission scenarios. Although the Cheetah scenario incurs 60% higher costs compared to the Hyena Scenario, it not only achieves the net-zero emissions target but also reduces the carcinogenic impact to 14% of that in the Hyena scenario. • AI-powered modeling integrates energy, emissions, and health to assess co-benefits. • PM2.5 predictions using XGBoost model for high accuracy with traffic and meteorological data. • Cancer risk assessment based on PM2.5 concentrations across socio-economic conditions.
Doraki et al. (Sun,) studied this question.