As the cyber security industry explores large language models (LLMs)--like GPT-4, PaLM, LaMDA, etc.--organizations are looking to define what return on investment these capabilities can bring for security programs. For cyber threat intelligence (CTI) functions, the increasing accessibility of LLMs shifts the interplay between the key components underlying any CTI program's ability to support its organization's threat intelligence needs. Organizations that look to apply LLM capabilities to CTI use-cases can expect categorically expanded capability to convert raw data into answers of key intelligence questions....
By: Ron Graf , John Miller
Full Abstract and Presentation Materials: www.blackhat.com/us-23/briefi...