A new ground-breaking cybersecurity tool has been launched that thinks like a human but works at the speed of a computer.
The machine learning AI device was developed and released by cyber defense company, Darktrace. It is revolutionary in digital defense because of the way that it has been trained to make human-like decisions at such high velocity.
The quickness at which cybersecurity judgments are made has been cut by 92% with the new design. Whereas it normally takes a human around 30 minutes to an hour to explore a single suspicious security report.
“Darktrace’s Cyber AI Analyst quickly presents security information in a format that’s both elegant and intuitive,” said Chris Kissel, research director at IDC. “By automatically investigating security events, the AI Analyst helps reduce noise more than any other technology. This is an important development in the security industry.”
The means of development
Using a combination of supervised and unsupervised deep learning, researchers took three years to foster the AI machine using algorithms that reflect on human thought processes and insights.
The company has already previously released an AI machine that detects cybersecurity issues. However, the newly announced Cyber AI Analyst investigates threats that use human ways of thinking that are actually built into the machine.
Darktrace took advantage of its previous design to develop the Cyber AI Analyst by combining the original qualities of its Enterprise Emmanual System with the newer research methods to make this tool. The AI then learned from “millions of interactions” between the techniques. This was part of what gives it both a human and a robotic way of thinking.
A help to humans
“The burden of investigating threats typically falls on the shoulders of a small number of trained security professionals,” said Mike Beck, global head of Threat Analysis, Darktrace. “With Cyber AI Analyst, security teams can now rely on AI to investigate hundreds of threats at once, allowing human analysts to focus on the most strategic work.”
100 of the world’s best cyber analysts were used in the research.