New Research: Enhancing Botnet Detection with AI using LLMs and Similarity Search | Rapid7 Blog

Rapid7 researchers used AI large language models (LLMs) to detect botnets leveraging TLS encryption. By analyzing embedding similarities in TLS certificates, the model achieved a 0.994 accuracy rate, outperforming proprietary alternatives in speed and cost-efficiency. This approach enables the detection of both known and zero-day botnets, reducing manual inspection and false positives for cybersecurity teams.