Researchers have developed a method for verifying the safety of AI-based collision avoidance systems in aviation. The approach involves assessing the representativeness and completeness of the Operational Design Domain (ODD) used during development and verification. This is crucial because the European Union Aviation Safety Agency requires that AI and Machine Learning systems demonstrate their ability to handle various scenarios and data distributions. The new method uses statistical distribution comparison methods, such as the Kullback-Leibler divergence and Cramér's V, to evaluate representativeness. It was tested using experimental data from previous collision avoidance system simulations. This work contributes to a systematic Safety-by-Design AI engineering process aligned with emerging EASA guidance. --- Why it matters: This matters because it provides a structured engineering process for defining target distributions and evaluating representativeness within ODDs, which is essential for ensuring the safety of AI-based systems in aviation. Source: https://arxiv.org/abs/2608.20864