A difference-in-difference approach to detecting AI-generated image

Diffusion models are able to produce AI-generated images that are almost indistinguishable from real ones. This raises concerns about their potential misuse and poses substantial challenges for detecting them. Many existing detectors rely on reconstruction error — the difference between the input image and its reconstructed version — as the basis for dis tinguishing real from fake images. However, these detectors become less effective as modern AI-generated images be comeincreasingly similar to real ones. To address this chal lenge, we propose a novel difference-in-difference method. Instead of directly using the reconstruction error (a first order difference), we compute the difference in reconstruc tion error– a second-order difference– for variance reduc tion and improving detection accuracy. Extensive exper iments demonstrate that our method achieves strong gen eralization performance, enabling reliable detection of AI generated images in the era of generative AI