Darktrace has added 70 new machine learning models and over 80 new features to its flagship AI cybersecurity platform.
The Cambridge-based company, which utilises self-learning AI to safeguard businesses across various industries, was created in 2013 by mathematicians and cyber defence professionals.
As part of Darktrace’s Antigena autonomous response technology, machine learning is utilised to make thousands of “micro-level” decisions in the background.
Antigena has been enhanced with 70 new machine learning models to boost its ability to neutralise threats automatically in real time.
“The hallmark of a great AI solution is the ability to surpass automation to seamlessly blend into users’ everyday work rhythm,” said Jack Stockdale OBE, CTO of Darktrace.
“When developing Darktrace Cyber AI products, our goal is to augment and uplift the security team to make the task at hand more efficient, so the end product is very intuitive and helps users in their workflow journeys.”
Darktrace has made a concerted attempt to follow the concepts of XAI (Explainable AI). XAI ensures that humans can access and comprehend the AI’s judgments.
Natural language processing is used in the incident display for Cyber AI Analyst to “clearly define the steps a human analyst would take if studying the same behavior and displays a succinct incident summary outlining each stage, which is easy to grasp and rapid to triage.”
In addition, any relevant events, such as associated users, destination ports, and protocols employed, will be highlighted in the incident display. A human analyst may delve into each incident reaction thanks to the detailed breakdown of activities executed by Darktrace’s solution.
Another key improvement in this release is to the Enterprise Immune System. Users can now use filters to narrow down incidents that have a particular severity or relate to specific classifications like compliance.
“With the latest release of Darktrace’s Enterprise Immune System, we really kept the user at the forefront of all UX/UI design decisions, from the beginning to the end of the AI product development life cycle,” explained Stockdale.