Devil in the Lens: Analyzing and Defending Physical Prompt Injection Against Vision-Language Models on Wearable Devices

• URL: http://arxiv.org/abs/2607.10269v1
• Date: Sat, 11 Jul 2026 11:56:05 GMT
• Title: Devil in the Lens: Analyzing and Defending Physical Prompt Injection Against Vision-Language Models on Wearable Devices
• Abstract summary: Vision-Language Models (VLMs) are rapidly deployed on human-facing wearable devices such as smart glasses to enable multimodal perception and AI-assisted decision-making.<n>Our work characterizes how malicious textual information embedded in physical environments introduces a high-priority visual channel for prompt injection.<n>Using physically captured photos from AI glasses in over 200 real-world environments, our analysis identifies 6 representative threat vectors of physically injected prompts.<n>Results show that these attacks consistently manipulate model outputs across integrity- and safety-critical tasks, achieving attack success rates of up to 96% and 60% in simulated and real-world settings.
• Score: 7.183439319638428
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision-Language Models (VLMs) are rapidly deployed on human-facing wearable devices such as smart glasses to enable multimodal perception and AI-assisted decision-making. While prior research has demonstrated the risks of visual prompt injection into digital image inputs of VLMs, the unique security challenges posed by the increasing integration between physical environments and wearable intelligence, such as those embodied in VLM-enabled AI glasses, remain underexplored. Toward understanding and modeling such threats, our work characterizes how malicious textual information embedded in physical environments introduces a high-priority visual channel for indirect prompt injection, where scene texts that hinder or evade human perception could hijack VLM models' behavior. Such \textit{Physical Prompt Injection Attacks} can not only disrupt normal tasks of VLM-enabled wearable devices, but also steer models to produce profane, biased, or even untruthful outputs. Using physically captured photos from AI glasses in over 200 real-world environments, our analysis identifies 6 representative threat vectors of physically injected prompts, and further evaluates their impacts on 12 VLM models. Results show that these attacks consistently manipulate model outputs across integrity- and safety-critical tasks, achieving attack success rates of up to 96\% and 60\% in simulated and real-world settings. Our analysis confirms that multiple models exhibit excessive blind trust in environmental text, ignoring the actual visual context and producing completely opposite summaries or directives. We further propose two targeted defense strategies, including a mask-based external filter and a semantic-vector-based internal detector, to effectively reduce the success rate and safety impact of these attacks.

Related papers

• Breaking the weakest link to evade vision language models [0.254890465057467]
  Vision Language Models (VLMs) have emerged as a critical component of multimodal AI systems.<n>In this work, we investigate the vulnerability of VLMs to adversarial perturbations applied to visual inputs.<n>We show that small, human-imperceptible perturbations can substantially alter the textual interpretation produced by the models.
  arXiv  Detail & Related papers  (2026-08-19T14:06:31Z)
• Towards Robustness against Typographic Attack with Training-free Concept Localization [42.575178574443946]
  Contrastive Language-Image Pretraining models (CLIP) serve as the foundational vision encoders for most modern Large Vision Language Models (LVLMs)<n>CLIP models exhibit a critical yet underexplored failure mode: irrelevant text appearing within images confounds visual representations, biasing them toward lexical meaning rather than true visual semantics.<n>This issue, commonly described as a Typographic Attack (TA), exposes a vulnerability that poses a significant risk to safety-critical applications such as autonomous driving.<n>We propose a novel, training-free mechanistic interpretability method to achieve interpretable and effective robustness against TA.
  arXiv  Detail & Related papers  (2026-07-02T17:55:24Z)
• Challenging Vision-Language Models with Physically Deployable Multimodal Semantic Lighting Attacks [23.938024446316717]
  Vision-Language Models (VLMs) have shown remarkable performance, yet their security remains insufficiently understood.<n>Existing adversarial studies focus almost exclusively on the digital setting, leaving physical-world threats largely unexplored.<n>We propose Multimodal Semantic Lighting Attacks (MSLA), the first physically deployable adversarial attack framework against VLMs.
  arXiv  Detail & Related papers  (2026-04-14T14:52:15Z)
• Universal Camouflage Attack on Vision-Language Models for Autonomous Driving [67.34987318443761]
  Visual language modeling for automated driving is emerging as a promising research direction.<n>VLM-AD remains vulnerable to serious security threats from adversarial attacks.<n>We propose the first Universal Camouflage Attack framework for VLM-AD.
  arXiv  Detail & Related papers  (2025-09-24T14:52:01Z)
• FreezeVLA: Action-Freezing Attacks against Vision-Language-Action Models [124.02734355214325]
  Vision-Language-Action (VLA) models are driving rapid progress in robotics.<n> adversarial images can "freeze" VLA models and cause them to ignore subsequent instructions.<n>FreezeVLA generates and evaluates action-freezing attacks via min-max bi-level optimization.
  arXiv  Detail & Related papers  (2025-09-24T08:15:28Z)
• Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents [54.35629963816521]
  This work introduces VIBMA, the first clean-text backdoor attack targeting VLM-based mobile agents.<n>The attack injects malicious behaviors into the model by modifying only the visual input.<n>We show that our attack achieves high success rates while preserving clean-task behavior.
  arXiv  Detail & Related papers  (2025-06-16T08:09:32Z)
• Natural Reflection Backdoor Attack on Vision Language Model for Autonomous Driving [55.96227460521096]
  Vision-Language Models (VLMs) have been integrated into autonomous driving systems to enhance reasoning capabilities.<n>We propose a natural reflection-based backdoor attack targeting VLM systems in autonomous driving scenarios.<n>Our findings uncover a new class of attacks that exploit the stringent real-time requirements of autonomous driving.
  arXiv  Detail & Related papers  (2025-05-09T20:28:17Z)
• Transferable Adversarial Attacks on Black-Box Vision-Language Models [63.22532779621001]
  adversarial attacks can transfer from open-source to proprietary black-box models in text-only and vision-only contexts.<n>We show that attackers can craft perturbations to induce specific attacker-chosen interpretations of visual information.<n>We discover that universal perturbations -- modifications applicable to a wide set of images -- can consistently induce these misinterpretations.
  arXiv  Detail & Related papers  (2025-05-02T06:51:11Z)
• Tit-for-Tat: Safeguarding Large Vision-Language Models Against Jailbreak Attacks via Adversarial Defense [90.71884758066042]
  Large vision-language models (LVLMs) introduce a unique vulnerability: susceptibility to malicious attacks via visual inputs.<n>We propose ESIII (Embedding Security Instructions Into Images), a novel methodology for transforming the visual space from a source of vulnerability into an active defense mechanism.
  arXiv  Detail & Related papers  (2025-03-14T17:39:45Z)
• Manipulation Facing Threats: Evaluating Physical Vulnerabilities in End-to-End Vision Language Action Models [79.60405008196139]
  Vision Language Action Models (VLAMs) are being proposed to achieve better performance in open-vocabulary scenarios for robotic manipulation tasks.<n>In this paper, we comprehensively evaluate VLAMs in the face of potential physical threats.
  arXiv  Detail & Related papers  (2024-09-20T03:02:05Z)
• Safeguarding Vision-Language Models Against Patched Visual Prompt Injectors [31.383591942592467]
  Vision-language models (VLMs) offer innovative ways to combine visual and textual data for enhanced understanding and interaction. Patch-based adversarial attack is considered the most realistic threat model in physical vision applications. We introduce SmoothVLM, a defense mechanism rooted in smoothing techniques, to protectVLMs from the threat of patched visual prompt injectors.
  arXiv  Detail & Related papers  (2024-05-17T04:19:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.