AI and Data: A Powerful Partnership to Transform Cybersecurity
Organizations expect more from AI and data than ever before. They want key insights, greater integration across the business, more control, and improved security. Now BlackBerry and Databricks are working together to create this experience and deliver a more powerful partnership between AI and data—one that can transform cybersecurity.
“We’re at a lifecycle refresh of cybersecurity architecture right now,” Jesse Scott told me. He’s Senior Director of GTM Cybersecurity at Databricks. “It’s a choice between vendor lock-in or a more open system that can grow with new challenges. We're big proponents of the latter and think Databricks can play a big role in that, working with BlackBerry to provide services on top.”
The Databricks platforms are intended to be open for this reason. “The thing that everyone's looking for is more control and oversight,” says Omar Khawajar, VP & CISO, Databricks. “They'd rather partner with someone to do that, but they still want to feel like they control and manage their own data.” Companies also want to be able grow as their threat surfaces and risks increase. “However, they don't want to feel like they have to make significant choices in their core data platform.”
You can watch my complete discussion with Omar Khawajar and Jesse Scott, which we had during RSA 2024, in the video below, or keep reading for excerpts.
BlackBerry and Databricks: Harnessing AI and Data
The strategic partnership between BlackBerry and Databricks is helping customers maintain control over their own data while combating cyberthreats. The two companies are working together on new initiatives based on the Databricks Mosaic AI platform. “At RSA, everything seemed to be about AI,” says Khawajar. “Databricks is an AI company, so we get excited when more people do AI because we want AI to be democratized.”
Lots of companies at the conference claimed they could defend against AI-powered attacks. “Unfortunately, they consider these attacks the same as regular attacks, so they don't think they need to do anything differently to protect against them.” Reality tells a different story, especially because of the increased pace of attack development uncovered in the latest BlackBerry Global Threat Intelligence Report.
AI to Improve Cybersecurity Posture and Threat Detection
AI can also be part of the solution. BlackBerry has been developing proven AI and machine learning for over a decade and is now seeing opportunities with partners like Databricks to extend these capabilities. This isn’t so much to stop AI-based attacks, but to solve security problems faster and more easily. BlackBerry recently announced the general availability of its Cylance® Assistant, a generative AI cybersecurity advisor that speeds up decision-making and stops more threats faster with fewer resources. It deploys AI to get unique insights out of data.
Cylance Assistant is designed to avoid the common pitfalls of early generative AI assistants, which can leak sensitive data, transgress privacy and IP theft regulations, and return misleading or false information. The Assistant is built directly into the Cylance® operations console and workflow rather than being standalone. This also facilitates feeding information into the LLM (large language model) through direct integration.
“Customers ask how they can use AI to improve their cybersecurity detection techniques, posture, and compliance,” says Scott. Customers also want to secure AI and make sure they are using it safely and securely. Databricks has released a white paper called The Databricks AI Security Framework that looks at some of the ways to do this. It also announced DBRX, its new foundational LLM, enabling organizations to leverage cost-effective generative AI trained on their own data.
“Most CISOs did not grow up in the world of data,” says Khawajar. “It's not a realm that they totally understand.” Databricks asked 100 CISOs to draw a picture of what they thought AI was. “They said they couldn’t. But if you can't draw a picture, do you know how to protect against it?”
Databricks divides AI into its constituent systems. “We can then explain the risks associated with each of those components and tell you what controls you need to deploy to mitigate each of those risks,” says Khawajar. “Now you've got something that you can operationalize, and you're not just saying you're securing AI as an aspiration. You have a framework and a methodology by which to make it happen in practice. We want our customers to get to these outcomes safely, securely, and responsibly.”
“Working for a data and AI company, I see there's a cultural shift needed because data and identity are big problems,” says Scott. “But it seems like there's not enough crosstalk between the data teams and security teams. They need to make that shift as quickly as possible to ensure they're ready for this generation.” Data and security teams don’t have to be best friends, but they help an organization if they are.
BlackBerry and Databricks: Looking Toward the Future
The long-term vision of the partnership between Databricks and BlackBerry is to guide this convergence of data, AI, and cybersecurity. Leading-edge companies see the need for centralized data. “Legacy architectures silo the security data, but so many other business units could derive value from it and vice versa. Security teams can derive value from other business data to find different attack vectors and anomalies.”
“Everyone needs to be pulling in data and to interact with it and manipulate it,” concludes Khawajar. As BlackBerry and Databricks look towards further collaborations AI will take an increasingly central role in keeping company data safe from cyberthreats. “It’s a journey that is democratizing data, leveraging AI, and finetuning models. In a few years, we will look back, and it will be like any other productivity tool.”
Watch the full video interview with Databricks from the RSA Conference, above.
