paperAugust 2026Unreviewed

The Boy Who Cried Wolf: Adversarial Misclassification of Safe Inputs as Unsafe in Multimodal Guardrails

Shuo Shi, Rui Yin, Naen Xu, Jiahao Chen, Chunyi Zhou, Tianyu Du, Zhihui Fu, Jun Wang, Zhaoxiang Wang, Shouling Ji

Abstract

Multimodal guard models have emerged as critical safety components for screening content in vision-language systems. While adversarial research has extensively studied jailbreaking attacks that produce false negatives, the inverse threat of inducing false positives on benign inputs remains unexplored. We introduce Unsafe Induction Attacks, where adversaries distribute imperceptibly perturbed safe images that trigger guard models to reject legitimate user requests, causing a "Boy Who Cried Wolf"

Framework mappings

OWASP Top 10 for LLM Applications

LLM01Prompt Injection

MITRE ATLAS

AML.T0054LLM Jailbreak

Suggested from the entry's categories.

Cite

@misc{shi2026boy,
  title = {{The Boy Who Cried Wolf: Adversarial Misclassification of Safe Inputs as Unsafe in Multimodal Guardrails}},
  author = {Shuo Shi and Rui Yin and Naen Xu and Jiahao Chen and Chunyi Zhou and Tianyu Du and Zhihui Fu and Jun Wang and Zhaoxiang Wang and Shouling Ji},
  year = {2026},
  month = aug,
  eprint = {2608.01373},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2608.01373}
}