DiSCO is a zero‑shot, black‑box defense for text‑to‑image models that operates solely at the prompt level. It expands prompts with a distribution‑guided suffix using beam search and contrastive scoring against safe and unsafe image pools generated by the target model, iteratively refining until safe content is produced. The method improves safety on the I2P benchmark under various red‑teaming attacks, reducing attack success rates by 37.7% and 25.13% while preserving semantic fidelity and image coherence.
By Tong Zhang, Motasem Alfarra, Carlos Hinojosa, Christos Louizos, Bernard Ghanem
arXiv:2604. 05853v3 Announce Type: replace Abstract: Modern text-to-image (T2I) models can now render legible, paragraph-length text, enabling a fundamentally new class of misuse.
By Zonghao Ying, Haowen Dai, Lianyu Hu, Zonglei Jing, Quanchen Zou, Yaodong Yang, Aishan Liu, Xianglong Liu
arXiv:2602. 10179v2 Announce Type: replace-cross Abstract: Recent advances in large image editing models have shifted the paradigm from text-driven instructions to vision-prompt editing, where user intent is inferred directly from visual inputs such as marks, arrows, and visual-text prompts.
By Jiacheng Hou, Yining Sun, Ruochong Jin, Haochen Han, Fangming Liu, Wai Kin Victor Chan, Alex Jinpeng Wang
The paper introduces the Detection Surface, a geometric framework that maps the decision boundaries of heterogeneous safety filters in text‑to‑image models. Using this insight, the authors propose CRACK, a multi‑agent debate system that iteratively mutates prompts, diagnoses layer‑specific constraints, and refines attacks to bypass composite defenses. Experiments demonstrate that CRACK can achieve attack success rates up to 99.63% while using fewer queries and preserving semantic fidelity.
By Kaiyan Wen, Shijie Zhang, Lu Yu, Guangdong Bai
The safety of large vision-language models is increasingly stress-tested by multimodal jailbreaks, yet existing attacks remain largely static at the meta level: template-based attacks freeze the image...
The paper introduces Meta-Adaptive Multimodal Jailbreaking (MAMJ), a method that jointly optimizes an attack strategy prompt and attacker weights to generate more effective jailbreaks against vision‑language models. Using an LLM‑based critique to refine the strategy and group‑level success‑rate rewards to update the weights, MAMJ achieves high attack success rates on MM‑SafetyBench, outperforming existing baselines by up to 24.1 percentage points. The learned attacker also transfers to unseen models and remains robust against typical defenses, highlighting a systemic vulnerability in current VLMs.
By Benlei Cui, Shen Pang, Yuke Wang, Xuemei Dong, Yuwen Zhai, Jingqun Tang, Haiyang Yu, Hui Xue, Longtao Huang, Haiwen Hong
Text-to-image diffusion models have achieved high visual fidelity and broad adoption, but remain vulnerable to safety violations when adversaries exploit them to synthesize illicit content. Existing alignment paradigms, from input sanitization to structural feature pruning, are largely organized around unsafe concepts explicitly exposed during filtering, editing, or localization.
The paper introduces TempJail, a temporal jailbreak framework targeting image‑to‑video generation models. It exploits a newly identified vulnerability where unsafe semantics arise from the composition of frames over time, rather than from single‑frame violations. By decomposing malicious captions into visual conditions and temporal instructions, and by employing controlled latent perturbations and template rewriting, TempJail achieves a 23.3 % higher attack success rate than prior methods on several commercial models.
By Qi Lu, Zehui Guo, David Yuanda Gan, Zijing Li, Hengda Zhang, Weijun Xu, Qiankun Zhang
Recent advancements in Image-to-Video (I2V) generation have transformed input images from simple appearance references into interactive control interfaces where visual cues such as arrows, sketches, and emojis orchestrate complex video dynamics with unprecedented controllability. However, these seemingly innocuous static cues can be interpreted by models as executable temporal instructions, unfolding into harmful actions in the generated videos.
The paper introduces NarrativeAttack, a jailbreak framework that exploits unified multimodal models (UMMs) by embedding a malicious query within a self‑contained three‑act visual narrative. The attack uses the model’s own image generator to create setup and resolution images, hiding the malicious event as a hidden climax, and concludes with an image‑based guessing game that forces the model to select the relevant answer. Experiments demonstrate that NarrativeAttack outperforms previous methods, achieving up to 88.25% attack success rate on Gemini‑2.5‑Flash, revealing a significant safety vulnerability in UMMs.
By Shaoxiong Guo, Tianyi Du, Lijun Li, Yuyao Wu, Jie Li, Jing Shao
The paper investigates how multimodal large language models are more susceptible to prompt injection when adversarial instructions are presented as text rather than as non-textual inputs like images. It proposes a training‑free defense that renders untrusted payloads into typographic images (or audio) before they reach the model, a method called Pictionary. Experiments on ten models and two benchmarks show that this approach significantly lowers attack success rates while maintaining normal functionality, and that fine‑tuning on image‑rendered instructions can further reduce the modality gap.
By Jie Zhang, Andrei Baroian, Jan N. van Rijn, Avital Shafran, Florian Tram\`{e}r
arXiv:2503.06989v5 Announce Type: replace-cross
Abstract: Recently, Multimodal Large Language Models (MLLMs) have demonstrated their superior ability in understanding multimodal content. However, the...
By Wenzhuo Xu, Zhipeng Wei, Xiongtao Sun, Zonghao Ying, Deyue Zhang, Dongdong Yang, Xiangzheng Zhang, Quanchen Zou