Jiahui Geng

Jiahui Geng

Assistant Professor at Linköping University. Research on trustworthy, efficient, and agentic AI.

Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities

Published in Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), 2025

This paper introduces a novel method for universal jailbreaking of multimodal large language models by exploiting non-textual modalities, revealing important security vulnerabilities in current multimodal AI systems.

Recommended citation: J Geng, TT Tran, P Nakov, I Gurevych. (2025). "Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities." Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics.
Download Paper