LLMs can create 10,000 malware variants evading detection with 88% success, degrading ML classifiers and risking AI model security.
AI Could Generate 10,000 Malware Variants, Evading Detection in 88% of Case
This was originally published on post