New Malware Embeds Prompt Injection to Evade AI Detection - Check Point Research
research.checkpoint.com
Detected for the first time, malware attempts AI evasion by injecting a prompt to tell the LLM to label the file as benign