Driving Scene Editing is the controlled modification of real or synthetic driving environments, including changes to weather conditions, lighting, object placement, road markings, and agent behaviors, to produce diverse training or evaluation data for autonomous driving systems. Generative models, including diffusion models and neural radiance fields, are increasingly used to apply edits while preserving scene geometry, temporal consistency, and photorealism. Research challenges include semantic coherence across edits, multi-level scene control, and scalable data generation pipelines.