Safeguarding LLMs Against Misuse and AI-Driven Malware Using Steganographic Canaries

Md Raz, Venkata Sai Charan Putrevu, Meet Udeshi +3

AI-powered malware increasingly exploits cloud-hosted generative-AI services and large language models (LLMs) as analysis engines for reconnaissance and code generation. Simultaneo…