At the Black Hat USA 2017 conference today, JASK unveiled an approach to applying artificial intelligence (AI) to IT security that employs crowdsourcing to continuously train the AI modules.
JASK CEO Greg Martin says JASK Trident is a security operations center (SOC) based on an instance hosted on Amazon Web Services (AWS) based on the open source TensorFlow machine learning software originally developed by Google. That software runs on top of an instance of the Apache Spark in-memory computing framework that is deployed on top of a distribution of Hadoop curated by Cloudera.
The most unique aspect of JASK Trident, says Martin, is that as a cloud service, it extends what it learns about IT security to all the customers that employ it. As each customer applies JASK Trident to its environments, whatever knowledge about a specific threat that is gained automatically gets shared with every other organization using the platform. Every additional new user of the platform also immediately benefits from all the previous training of the AI model by the JASK ecosystem.
“It’s pretty clear there’s a need for a new approach to security,” says Martin. “Organizations need to be willing to collaborate.”
Martin says that when it comes IT security, more organizations are interested in applying AI, largely because the shortage of IT professionals with security expertise is so chronic. Rather than employing those limited resources to perform the same rote tasks everyday, AI platforms such as JASK Trident make it feasible to apply what limited IT security expertise there is to identifying complex threats.
The sad truth of the matter is that existing approaches to IT security are not up to the task. It’s hard to say with any real conviction that the IT security war is being won at any level. There’s no silver bullet when it comes to IT security. But the level of relatively common cyberattacks that continue to get past existing defenses should make it painfully obvious that many aspects of IT security eventually will be done much better by a machine than a person.