When Generative Augmentation Hurts: A Benchmark Study of GAN and Diffusion Models for Bias Correction in AI Classification Systems

ArXi:2603.16134v1 Announce Type: cross Generative models are widely used to compensate for class imbalance in pipelines, yet their failure modes under low-data conditions are poorly understood. This paper reports a controlled benchmark comparing three augmentation strategies applied to a fine-grained animal classification task: traditional transforms, FastGAN, and Stable Diffusion 1.5 fine-tuned with Low-Rank Adaptation (LoRA). Using the Oxford-IIIT Pet Dataset with eight artificially underrepresented breeds, we find that FastGAN augmentation does not merely underperform at very low.