PALO Alto Networks' report discusses the state of AI-enabled malware, highlighting that while AI integration in malware is real, its actual deployment in the wild is limited. The study analyzed 405 malware samples, revealing that only 12 samples were detected on protected endpoints, reflecting that 97% of the samples exist only in research and sandbox environments.
Key findings include: AI-enabled malware mostly consists of proof-of-concept samples, security testing tools, and AI-themed brand abuse; existing security mechanisms effectively detect these threats without needing novel approaches; and while AI aids in developing malware, it hasn’t significantly improved operational deployment. The report concludes with recommendations for defenders to utilize existing detection measures to counteract evolving AI-enabled malware threats.