RA-Det: Towards Universal Detection of AI-Generated Images via Robustness Asymmetry

Xinchang Wang, Yunhao Chen, Yuechen Zhang, Congcong Bian, Zihao Guo, Xingjun Ma, Hui Li
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:125515-125532, 2026.

Abstract

Recent image generators produce photo-realistic content that undermines the reliability of downstream recognition systems. As visual appearance cues become less pronounced, appearance-driven detectors that rely on forensic cues or high-level representations lose stability. This motivates a shift from appearance to behavior, focusing on how images respond to controlled perturbations rather than how they look. In this work, we identify a simple and universal behavioral signal. Natural images preserve stable semantic representations under small, structured perturbations, whereas generated images exhibit markedly larger feature drift. We refer to this phenomenon as robustness asymmetry and provide a theoretical analysis that establishes a lower bound connecting this asymmetry to memorization tendencies in generative models, explaining its prevalence across architectures. Building on this insight, we introduce Robustness Asymmetry Detection (RA-Det), a behavior-driven detection framework that converts robustness asymmetry into a reliable decision signal. Evaluated across 14 diverse generative models and against more than 10 strong detectors, RA-Det achieves superior performance, improving the average performance by 12.92%. The method is data- and model-agnostic, requires no generator fingerprints, and transfers across unseen generators. Together, these results indicate that robustness asymmetry is a stable, general cue for synthetic-image detection and that carefully designed probing can turn this cue into a practical, universal detector.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-wang26ad, title = {{RA}-Det: Towards Universal Detection of {AI}-Generated Images via Robustness Asymmetry}, author = {Wang, Xinchang and Chen, Yunhao and Zhang, Yuechen and Bian, Congcong and Guo, Zihao and Ma, Xingjun and Li, Hui}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {125515--125532}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wang26ad/wang26ad.pdf}, url = {https://proceedings.mlr.press/v306/wang26ad.html}, abstract = {Recent image generators produce photo-realistic content that undermines the reliability of downstream recognition systems. As visual appearance cues become less pronounced, appearance-driven detectors that rely on forensic cues or high-level representations lose stability. This motivates a shift from appearance to behavior, focusing on how images respond to controlled perturbations rather than how they look. In this work, we identify a simple and universal behavioral signal. Natural images preserve stable semantic representations under small, structured perturbations, whereas generated images exhibit markedly larger feature drift. We refer to this phenomenon as robustness asymmetry and provide a theoretical analysis that establishes a lower bound connecting this asymmetry to memorization tendencies in generative models, explaining its prevalence across architectures. Building on this insight, we introduce Robustness Asymmetry Detection (RA-Det), a behavior-driven detection framework that converts robustness asymmetry into a reliable decision signal. Evaluated across 14 diverse generative models and against more than 10 strong detectors, RA-Det achieves superior performance, improving the average performance by 12.92%. The method is data- and model-agnostic, requires no generator fingerprints, and transfers across unseen generators. Together, these results indicate that robustness asymmetry is a stable, general cue for synthetic-image detection and that carefully designed probing can turn this cue into a practical, universal detector.} }
Endnote
%0 Conference Paper %T RA-Det: Towards Universal Detection of AI-Generated Images via Robustness Asymmetry %A Xinchang Wang %A Yunhao Chen %A Yuechen Zhang %A Congcong Bian %A Zihao Guo %A Xingjun Ma %A Hui Li %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wang26ad %I PMLR %P 125515--125532 %U https://proceedings.mlr.press/v306/wang26ad.html %V 306 %X Recent image generators produce photo-realistic content that undermines the reliability of downstream recognition systems. As visual appearance cues become less pronounced, appearance-driven detectors that rely on forensic cues or high-level representations lose stability. This motivates a shift from appearance to behavior, focusing on how images respond to controlled perturbations rather than how they look. In this work, we identify a simple and universal behavioral signal. Natural images preserve stable semantic representations under small, structured perturbations, whereas generated images exhibit markedly larger feature drift. We refer to this phenomenon as robustness asymmetry and provide a theoretical analysis that establishes a lower bound connecting this asymmetry to memorization tendencies in generative models, explaining its prevalence across architectures. Building on this insight, we introduce Robustness Asymmetry Detection (RA-Det), a behavior-driven detection framework that converts robustness asymmetry into a reliable decision signal. Evaluated across 14 diverse generative models and against more than 10 strong detectors, RA-Det achieves superior performance, improving the average performance by 12.92%. The method is data- and model-agnostic, requires no generator fingerprints, and transfers across unseen generators. Together, these results indicate that robustness asymmetry is a stable, general cue for synthetic-image detection and that carefully designed probing can turn this cue into a practical, universal detector.
APA
Wang, X., Chen, Y., Zhang, Y., Bian, C., Guo, Z., Ma, X. & Li, H.. (2026). RA-Det: Towards Universal Detection of AI-Generated Images via Robustness Asymmetry. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:125515-125532 Available from https://proceedings.mlr.press/v306/wang26ad.html.

Related Material