AI To Play A Decisive Role In Cyber Security

by Sohini Bagchi    Jan 08, 2018


The use of Artificial Intelligence is increasingly becoming commonplace and can be applied in every sphere of business for bigger gain. Given the state of cyber security today, the implementation of AI systems into this domain be a real turning point, believe experts. In fact, new generations of malware and cyber-attacks that are difficult to detect with conventional cybersecurity protocols and require more dynamic approaches can rely on AI or ML and respond better to such complex threats. Sunil Gupta, President and Chief Operating officer Paladion Networks, and a veteran in this area discusses the complex yet very interesting role of AI in cyber security and risk management. Excerpt.

CXOToday: As businesses increasingly become digital, are we at risk of losing control over the machines we are creating?

Gupta: My answer would be both ‘yes’ and ‘no. However, we can no longer cling to old definitions of control. The enterprise technology ecosystem is now much more complex, and much more open, than it used to be. We can no longer control our perimeters, because those perimeters no longer exist. At the same time, attackers have become much more creative and sophisticated. We can no longer control a set list of known attacks, because most attacks no longer fit recognizable patterns and signatures. On the other hand, that simply means we need to develop new definitions of control. Instead of attempting to control our perimeter against known attacks, we must learn to control unique attacks that have already breached our systems. There is still a place for prevention in cybersecurity. But we must extend our definitions of control to consider the full lifecycle of a threat with full Managed Detection and Response capabilities, which can discover and mitigate threats before catastrophic damages.

CXOToday: Does AI mean the end of human element in cyber security or they both can co-exist?

Gupta: They both must co-exist. There is currently a misconception about AI’s role in cybersecurity. Many people hear the term “AI” and imagine a general AI system that can think as creatively as a human—only better. This general AI might be developed in the future, but at the moment AI’s role in cybersecurity is very specialized. AI is good at processing a massive volume of threat data and providing a summary of what it found. But we still need humans to tell the AI what to look for, to review the AI’s outputs, and to ultimately decide how to act upon the AI’s findings. The goal of any AI-driven cybersecurity service is simple: Let AI do what it does best, and let humans do what they do best. That is why, at Paladion, we combine a cutting-edge AI-driven platform with the industry’s largest team of cybersecurity experts.

CXOToday: Can hackers not enter the machine and alter the path of AI?

Gupta: There is one sobering truth about cybersecurity: it doesn’t matter what defenses you put in front of attackers, they will eventually find a way around it. There is no “silver bullet” technology that will solve the problem of cybersecurity once and for all. Cybersecurity will always be a back-and-forth game between attackers and defenders that will constantly evolve as technology evolves. No matter how powerful AI is, it will not solve this problem. It is the next necessary step we need to take to secure our systems, but it is possible that attackers may one day find a way to bypass, or subvert, AI—the same way attackers have bypassed, or subverted, every other security measure cybersecurity experts have developed. Although It’s important to ask this question, it is moot a point today. Attackers are already using AI to power their attacks, and we must deploy AI-driven defenses to keep up. In the future, we could be looking at dynamic machine learning algorithms that morph once it senses an intrusion; without affecting the original purpose for which it was designed.

CXOToday: Paladion offers AI-driven managed security. How is it different from the ones available in the market?

Gupta: When you take an honest look at the market right now, you see that most AI-driven managed security is not particularly mature. At the moment, most cybersecurity providers are still MSSPs, or are MDR providers who only provide end point detection and response (EDR) or EDR with User and Entity Behavior Analytics (UEBA) . Paladion goes far beyond this to include four distinct types of analytics; EDR, UBA, Application Threat Analytics, and Network Threat analytics on a Big Data AI platform. “MDR” and “AI” are very hot topics right now, and clients are demanding them. So, we are seeing many MSSPs adopt the bare-minimum level of managed service and AI-integration to be able to state that they offer these services. But there is a significant difference between a last-generation security provider who has attempted to layer on next-generation security services, and a provider like Paladion that has built a proprietary AI-driven MDR service unlike any other in the world. We have stayed ahead of the curve, something I can confidently say.

CXOToday: What is the future for AI in Cyber Security?

Gupta: In the short-term, AI will become a standard element of cybersecurity as security providers rush to both meet client demand and to keep up with AI-driven crybercriminals. We recommend a thorough evaluation of AI claims before subscribing to AI driven cyber defense. In the long-term, AI will only play a greater role in cybersecurity. Yes, “AI” is a buzzword right now. You might worry it’s simply a fad. But when you look at the root causes driving the need for AI in cybersecurity, you will see that AI is a response to a fundamental shift in the threat landscape.

As enterprises further adopt the Cloud, Mobile, and IoT, they will open more vulnerability points for cybercriminals to exploit. As they further digitize, they will create opportunities for increasingly complex multi-channel attacks. Enterprise security staff are already flooded with more daily threat data than any human-only team could ever handle. This data will only increase every year, and so will the need for AI to effectively process it.