Wealth & Finance International - January 2017

Wealth & Finance International 44 Cyber - Challenging Conventional Wisdom ur award-winning software, originally commissioned by GCHQ and the MOD, uses machine learning to analyse web traffic; pinpointing high risk attacks to prevent data loss and keep businesses secure. Barely a day goes by where a high profile cyber-attack is not in the news. The average cost of a security breach is estimated to be around £4M and increasing, as attackers use the latest tools and techniques to launch sophisticated attacks. At the same time, software automation and artificial intelligence is being used as a weapon, exponentially increasing the volume of attacks and confounding security teams. Everybody (and everything!) is now a po- tential target, as attackers search relentlessly for weaknesses to exploit. The rewards continue to heavily outweigh the risks For the attackers, the risk vs reward is a simple equation. Global cybercrime is estimated to be costing around $400 billion annually. A teenager can hack into a corporate network with a few clicks of a commercially available tool. The chances of being caught are low. The chances of be punished are even lower, as law enforcement struggles to convict attackers operating across international jurisdictions. There is no doubt that cyber criminals are winning. A proportion of the profits is ploughed back into research and development of new offensive methods. For example, it is estimated that a new malware signature is created every 5 seconds. Polymorphic attacks that modify and adapt known attack signatures will always defeat reactive, rule-based defences. The conventional cyber defences that use rules to detect precise attack signatures were not designed to prevent these types of attacks. How- ever, most businesses continue to rely on this technology. Detecting an attack using a rule depends on something already having been attacked in that way. This is an archaic and reactive approach that is not effective and will continue to lead to businesses being compromised. As a result, public and corporate perception in defensive cyber security is at an all-time low. Something must change. It’s time to get smart! My co-founder and I have worked across UK security agencies and GCHQ, where we have delivered cyber security projects with particular expertise in defining and measuring information security risk. Cutting a long-story short, we were awarded contracts with GCHQ and the Min- istry of Defence (MOD) to research new ways of prioritising the growing number of attacks by using artificial intelligence techniques; i.e. using machines to cut through the deluge of data facing security analysts, to pinpoint the attacks that would cause harm. The research projects were successful and as a result we founded Cyberlytic in 2013. Our founding belief is that applying artificial intelligence within the cyber security domain, should enable security teams to be more efficient and reduce the demand on human opera- tors; in other words, it’s time to let machines defend against machine led attacks and allow our limited human cyber resources to focus on business-critical responses. This approach is revolutionising how cyber security is conducted. It is why we are seeing an influx of start-ups challenging the more estab- lished players. The scene is set for a few of these companies to take the initiative and effectively lead the battle against the attackers. At Cyberlytic, we have developed and patented an approach that doesn’t rely on rules to detect cyber-attacks. We use machine learning to determine and build the characteristics of an attack. Our software analyses web traffic and when it spots something that looks dangerous, it conducts an analysis. Essentially, our software has two core functions; detect the attack and determine if it poses a risk to the business. For example, it can determine if the attack is performing a reconnaissance (such as an automated tool), which is lower risk than an attack that enumerated a database. This analysis happens instantly. Our software even strips out the malicious part of an attack meaning they are stopped before they do any harm. All without maintaining a set of rules that are out of date the moment they are created. Looking ahead and making a change This is a huge year for Cyberlytic. We’ve deployed the software with large and small businesses and have had compelling results. Frequently our software detects and helps prevent attacks that bypass conventional firewalls. Due to these successful deployments, we’ve raised early-stage investment funding and are growing our data science and development team in Belfast, as well as our commercial teams in London and the USA. Cyberlytic is a world leading web application threat defence company. It provides cyber security software that detects, prioritises and prevents the most sophisticated web-based attacks. The firm’s CEO Stuart Laidlaw lifts the lid on his firm’s award-winning software and the impressive array of clients they work with. O 1701WF44