Using AI and machine learning to improve cyber defences and reduce human cyber risk

ai and machine learning blog image


We are CybSafe. We’re a British cyber security and data analytics company.

May 21, 2019



To understand how AI and machine learning can reduce cyber risk, it’s worth considering how a typical taxi journey will look in a few years time.

Today, when you jump into a taxi, you’re greeted by a driver who might inquire into your preference of radio station and ask you one or two cursory questions while edging you towards your eventual destination.

In a few years time, your journeys will be much the same – but with one key difference.

In a few years time, taxis are probably going to incorporate AI… and thus won’t include a driver at all.

How AI and machine learning work

Driverless cars are a nice demonstration of what AI and machine learning are and how they work. Quite simply, they consider the external environment and react in an appropriate manner.

In the case of a driverless taxi journey, computers might consider current vehicle speed and how far it is to an upcoming junction before adjusting the vehicle’s speed accordingly.

This, of course, is precisely what humans do (almost entirely subconsciously) when making decisions:

We consider “data”.

Then, we make decisions.


AI and machine learning in cyber security

Jumping back to the field of cyber security, AI and machine learning are already being leveraged to:

  • Improve the efficacy of spam filters
  • Detect the presence of malware in networks
  • Authenticate users
  • Prevent fraud

and pursue a whole host of other worthwhile goals.

How prominent are AI and machine learning in security today?

According to the Ponemon Institute’s fourth annual report on cyber resilient organisations, 23% of organisations are already using security technologies that “augment or replace human intervention in the identification and containment of cyber exploits or breaches” extensively.

It’s tempting to think if security-based AI and machine learning as something for the future. In reality, security teams have already started improving their defences with AI and machine learning – and to great effect:

According to the Ponemon Institute report, organisations that have embraced security automation are less likely to suffer breaches, less likely to experience significant disruption following a breach and are better able to prevent, detect, contain and respond to cyber attacks.


Using AI and machine learning to reduce human cyber risk

As well as using AI and machine learning as a technological defence, both can be used to address the human aspect of cyber security and reduce human cyber risk. How?

AI and machine learning work by considering data and reacting accordingly. CybSafe, the world’s first truly intelligent cyber security awareness, behaviour and culture solution that demonstrably reduces human cyber risk, continuously monitors vast quantities of data in three broad categories:

  • What individuals know about cyber security (ie, security awareness)
  • Individual security habits (ie, security behaviours)
  • And how people think and feel about security (ie, security culture)

CybSafe intelligently aggregates its findings to calculate your human cyber risk. It then uses what it “learns” about security awareness, behaviours and culture to launch appropriate security interventions.

The solution might learn, for example, that a particular team has a knowledge gap when it comes to secure browsing – and offer the people in the team personalised security awareness training as a result.

Or it might learn that a subset of individuals are more likely to click on simulated phishing campaigns that use authority to instil fear. Again, CybSafe deploys tailored security interventions as a result.

Following security interventions, CybSafe once again calculates human cyber risk. It repeats successes and, over time, CybSafe’s in-built AI-machine learning demonstrably reduces human cyber risk.


Why are so few organisations using AI and machine learning to reduce human cyber risk?

It’s interesting to note that, while many organisations now use AI and machine learning as a technological cyber defence, only a select few are using AI and machine learning to reduce their human cyber risk.

It’s difficult to say why this might be but it’s possible that, as AI and machine learning are technological innovations, security leaders are applying them as technological defences without thinking about how AI and machine learning could address the human aspect of cyber security.

Focusing on bolstering our technological defences only is a trap the security industry has fallen into before – and it partly explains why so many breaches today involve some form of human error.

Let’s hope that, this time around, we do what we’re programming computers to do.

Let’s hope that, this time around, we learn from the past.

Try it yourself or see it in action