Request a call

Cyber Security Disruption and Machine Learning in 2018

Last updated on February 7, 2018 by Dotsquares

Noting the notorious Equifax, WannaCry, Petya, and Eternal Blue cyber-attacks, the BBC has titled 2017 as the year of ‘Cyber-Geddon’. Its attacks like these that have revealed just how vulnerable we actually are in this digital era. What’s more is these vulnerabilities are likely to increase in 2018, due to the increased use of IoT and cloud services, with more concerns over data privacy, extortion, and ransom.

All these factors have contributed in the exponential increase in the cyber security market, which according to a report published by independent global market research firm, “Markets and Markets”, is to grow by the compound annual rate of 11% to become $232 billion by 2022, from $137.8 billion in 2017.

Potential Cyber Security threats

One of the biggest cyber security threats will be down to the increased application of cloud services. The increased usage of IaaS, PaaS, SaaS as well as computing platforms like Chromebook will definitely disrupt the way end-points are being handled in the more traditional cyber security practices. It is likely that the attackers will catch on quick to this change faster than the threat hunters and breach responders can.

The same problem may arise with the increased application of IoT devices. In fact, it’s easier for the attackers to tap into an IoT network by simply purchasing botnet kits from the dark web, than it is for the investigators to track the malicious activities and even to recover from Distributed Denial of Service. Another major threat, Ransomware, isn’t going anywhere either. If anything, ransomware attacks will most likely create an even stronger impact this year with more susceptible points, and tempting ransom available to the attackers than ever before.

As it stands technology is advancing at an exponential pace and cyber security is simply unable to catch up with it.

The Disruption

When it comes to technical advancements, one cannot ignore the progress made by AI and ML in the previous years. We have already witnessed how the Google’s Auto ML has created its “AI Child” NASNet, and how it was far better to other similar systems developed by humans. It projects that the present gap between the cyber security issues and solutions may as well find a more reliable answer with ML.

Actually, Machine Learning solutions in Cyber Security makes more sense when we realize that it is not only the potential threats in the system, but also for the ones present now. Brian Beyer, CEO of a Cyber security solution providing organization,“Red Canary”, has revealed in a recent interview with Julian Mitchell that only as much as 5% of alarms being raised by present security systems, are actually getting investigated. This severe shortfall can be accounted to the extreme gap between the demand and supply of the required experts. There are simply not enough resources with the required skill sets to deal with these problems.

This is where Machine Learning comes into the picture. As Brian mentions, his team simply installs tools in the end-point devices of the secured system. This tool then assesses the behavioural changes and data to determine whether the person accessing the system is authorized or not.

This approach is a tad different than that of traditional tools that try to capture certain suspected activities, like encryption of data, to raise alarms. While these traditional methods are effective they are not impregnable. Besides, implementation of similar ML-powered tools can really intensify the detection, as well as the recovery processes against potential security breaches.