Darktrace further builds out Cyber AI Loop with launch of Darktrace PREVENT

Darktrace PREVENT complements the existing DETECT and RESPOND solutions, and will integrate with the final component, HEAL, when it comes out next year.

David Masson, director of Enterprise Security for Darktrace

Cybersecurity vendor Darktrace has announced the launch of Darktrace PREVENT, the third of the fourth modules in the company’s integrated Cyber AI Loop capabilities, an interconnected set of AI products that deliver a proactive cyber security capability to help organizations pre-empt future cyber-attacks. Darktrace PREVENT is also the third product area in Darktrace’s delivery of a Cyber AI Loop, the industry-first set of AI capabilities which work together autonomously to optimize an organization’s state of security through a continuous feedback loop.

“We’ve been doing DETECT since 2013 and RESPOND since 2017, but until two weeks ago, we were only using AI to stop attackers from causing damage,” said David Masson, director of Enterprise Security for Darktrace. “Now we will use AI to prevent attacks from happening in the first place.”

The new Darktrace PREVENT product family is based on breakthroughs developed in the firm’s Cambridge Cyber AI Research Centre and the capabilities gained through the acquisition of Cybersprint in March 2022. PREVENT uses AI to ‘think like an attacker,’ finding pathways to an organization’s most critical assets from inside and outside.

“This ability of PREVENT to look at both our inside and outside views and send them to DETECT and RESPOND is a differentiator for us,” Masson said. While the original focus when they started working on PREVENT about two years ago was inside, Masson indicated that they have now expanded it to find vulnerabilities outside the organization.

Masson also acknowledged that more than a few products do something similar, but stressed that there is a fundamental difference between them.

“Others use supervised machine learning – training the machine to do something in the hope that it will guess what the next bad stuff will be,” he indicated. “A machine will only find what it has been trained to find. With our unsupervised machine learning, we don’t train it. The machine learns how the business works from the ground up, and looks for change, which it can do because the AI is very clever. Other products tell you what happened. We focus on what’s happening and how we can stop damage. That’s a fundamental difference between us and most of the companies in the market. Because we understand an organization in its entirety, and because PREVENT has that understanding, it can show you pathways to your most critical assets, and show you an attack path before the bad guys find it.”

Masson expects that the market for PREVENT will be fairly broad.

“It’s really for everybody with digital infrastructure they need to protect,” he said.

Because the inside and outside components of PREVENT will be sold as two separate products, this provides an additional benefit for channel partners, he added.

“It will give them two new products to sell,” he said.

While DETECT and RESPOND are both available now, PREVENT is scheduled to be available on August 1, 2022.

HEAL, the final component scheduled at this time within the Cyber AI Loop, is scheduled to be available next year.