According to the World Health Organization, nearly 90% of the global population is exposed to polluted air due to rapid industrialization, urbanization, and other modern developments that have severely deteriorated air quality. In 2023 alone, outdoor air pollution accounted for approximately 8.34 million deaths worldwide, contributing to chronic illnesses such as respiratory disorders, heart disease, asthma, and lung cancer. To address these growing concerns, this study employs artificial intelligence and machine learning techniques to analyze air quality data from various Indian cities collected between 2015 and 2020 by the Central Pollution Control Board of India. Through data preprocessing, statistical analysis, and the use of explainable AI, the study identifies PM2.5, PM10, and CO as the most hazardous pollutants contributing to severe Air Quality Index (AQI) levels in metropolitan regions. Among the evaluated models, the Random Forest regressor demonstrated superior performance with an R2 score of 96.17% on test data. The proposed framework provides valuable insights that can assist researchers and policymakers in formulating effective strategies to mitigate air pollution and enhance public health outcomes.