Metraipolis › archive › dc_ba6cd39e2d62

One Attack to Fool Them All: Highly Transferable Black-Box Adversarial Attacks on Frontier MLLMs — arXiv:2609.33833

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33833
License (as archived)
CC0-1.0
Archived
2026-10-07T22:40:13.557Z
SHA-256 of the archived text
de6e6117d8ebf12d71a0e39b33f528374b6c8d225823f04847ac926fbe5b2c5a
Stable URL of this page
https://metraipolis.com/documento/dc_ba6cd39e2d62

How to cite

One Attack to Fool Them All: Highly Transferable Black-Box Adversarial Attacks on Frontier MLLMs — arXiv:2609.33833 — arXiv (en), https://arxiv.org/abs/2609.33833 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ba6cd39e2d62 (Metraipolis archive dc_ba6cd39e2d62, sha256 de6e6117d8ebf12d71a0e39b33f528374b6c8d225823f04847ac926fbe5b2c5a)
BibTeX
@misc{dcba6cd39e2d62,
  title = {One Attack to Fool Them All: Highly Transferable Black-Box Adversarial Attacks on Frontier MLLMs — arXiv:2609.33833},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33833},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ba6cd39e2d62; sha256 de6e6117d8ebf12d71a0e39b33f528374b6c8d225823f04847ac926fbe5b2c5a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33833. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Adversarial attacks have long posed a fundamental threat to machine learning systems. As multimodal large language models (MLLMs) rapidly evolve and become widely deployed, assessing their vulnerability to such attacks is essential for their safe use. In this work, we investigate whether a single adversarial image can consistently mislead diverse frontier MLLMs in black-box settings. We propose O-Attack, a highly transferable black-box attack framework. This framework builds on our insight that surrogate models contain a broad, high-level, cross-modally aligned semantic space. This space extends beyond final-layer outputs and provides multiple semantically consistent representations that remain underexploited by existing attacks. Within this space, O-Attack anchors aligned representations, progressively broadens semantic conditions, and optimizes perturbations through semantic consensus to promote consistent target alignment. By fully exploiting this space with the same surrogate models as M-Attack, O-Attack raises attack success rates on GPT-5.4 (29.1% to 77.2%), Claude-4.6 (42.8% to 81.6%), and Gemini-3.1 (38.2% to 80.9%). Extensive experiments across 24 MLLMs show that O-Attack outperforms six state-of-the-art methods in black-box transferability, with consistent effectiveness across prompts and improved efficiency and imperceptibility. This work exposes the practical safety risks posed by black-box adversarial attacks against frontier MLLMs, underscoring the need for more rigorous robustness evaluation and more effective defenses.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.33833 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T22:14:28.783Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).