Artificial Intelligence To Crawl DeepWeb To Find Cyber Attacks Before They Happen

A new joint operation undertaken by military contractor Lockheed Martin, along with Cyber Reconnaissance, Universidad Nacional del Sur and the Universities of Arizona and Southern California aims to revolutionize cyber intelligence going forward throughout the future. The conception is not dissimilar to the ‘Web Bot” project, which attempts to index/collect, analyze and interpret key words and searches sent across the internet in live time as they are happening in order to predict future trends.

The project is officially called DARKMENTION, “a deployed system to predict enterprise-targeted external cyber attacks.” It was developed under the Cyberattack Automated Unconventional Sensor Environment (CAUSE) program for the Intelligence Advanced Research Projects Activity (IARPA), and attached to a database belonging to CYR3CON, a threat intelligence firm, which has cataloged various successful cyber attacks/attack styles pulled off throughout the past.

Together along with this database, DARKMENTION aims to crawl both the “Surface Web” and “Deep Web” in hopes of finding, recording and indexing various hacking forums, website and chats, in order to pick up on key issues or talking points within them. In doing so, researchers hope to uncover potential targets before hackers can organize enough to launch attacks against them, and/or uncover new information about how attacks are are actually set up/pulled off. The project could theoretically revolutionize intelligence gathering, while also drastically cutting down on Governmental/military operational budgets.

If this seems impossible to you, you might remember a 2016 study by Intelliagg and Darksum, which managed to crawl, index and catalog approximately 29,532 websites hosted on the “DarkNet” – roughly 54% of the entire estimated size of the DarkNet. I bring this up because Darksum‘s project didn’t have nearly the budget or resources as Lockheed Martin‘s now does, and the Surface Web and Deep Web are inherently much easier to navigate than the DarkNet.

Full Research Project: