Causal AI in the Automotive Industry: Impact Analysis Through Carbon Emission Case Study

Artificial Intelligence (AI) is increasingly used across various industries, but a significant limitation is its reliance on correlations rather than understanding cause-and-effect relationships. This correlation-based approach often results in “black-box” models, where the reasoning behind decisions is unclear. Causal AI offers a promising solution by identifying and quantifying cause-and-effect relationships within a system. This approach involves two main steps: causal discovery, which reveals potential causal links in a dataset, and causal inference, which measures the strength of these relationships. In this research, we explored the application of causal AI in the automotive industry to improve sustainability, a crucial topic given the significant environmental impact of vehicle emissions. We applied various causal AI algorithms for comparison and conducted a case study to demonstrate our methodology. The selected dataset was well-described by existing models, allowing us to validate our results against established benchmarks. While our study primarily focuses on key emission factors, the methodology is extendable to include complex variables, such as human behavior, for which no current models exist. Additionally, we made the source code publicly available to support further research.