skip to content
Decrypt LOL

Get Cyber-Smart in Just 5 Minutes a Week

Decrypt delivers quick and insightful updates on cybersecurity. No spam, no data sharing—just the info you need to stay secure.

Read the latest edition

LLMs Improve Evasion Techniques for Malicious JavaScript

/ 1 min read

🌀 LLMs Enhance Malware Evasion Techniques in JavaScript. Researchers have developed an adversarial machine learning algorithm that utilizes large language models (LLMs) to generate novel variants of malicious JavaScript code, improving detection rates of such threats by 10%. While LLMs struggle to create malware from scratch, they can effectively rewrite existing malicious code, making it harder for detection tools to identify. The algorithm employs a series of transformations, such as variable renaming and dead code insertion, to produce thousands of unique malware variants that maintain their malicious functionality. In response, the team retrained their detection model on these LLM-rewritten samples, significantly enhancing its ability to identify evolving threats in real-time. This ongoing research aims to stay ahead of emerging AI-driven cyber threats.

Source
{entry.data.source.title}
Original