1 paper
Bhavik Chandna, Mariam Aboujenane, Usman Naseem
Large Multimodal Models (LMMs) are increasingly vulnerable to AI-generated extremist content, including photorealistic images and text, which can be used to bypass safety mechanism…