Today’s world is a very different one to that of just 30 years ago. Where we previously used to be terrified of the idea of one of the superpowers using atomic bombs, today we are scared of somebody hacking our computer systems and destroying our infrastructures. It is a very real fear and one that, unlike atomic bombs, has been regularly unleashed on people.
However, we are not just sitting ducks in the face of this increasing threat and, behind the scenes, work is constantly being done to increase the protection that companies and individuals have from nefarious hackers. One of the most powerful tools they have in this constantly evolving battle is in the use of data.
One of the most progressive industries in this regard is banking and financial services, where a huge amount of money combined with a high number of transactions has meant that it is the perfect environment for data analytics to have a big impact. It is, unfortunately, something desperately needed as we have seen significant increases in the amount of card fraud. In the UK, online banking fraud increased 26% between 2013 and 2014, for instance. In a country of only 63 million people, there were 53,192 cases of online banking fraud where people incurred loss of money. The numbers in the US are even more shocking, with the country responsible of 47% of all credit card fraud despite accounting for only 24% of total card transactions (http://www.creditcards.com/credit-card-news/credit-card-security-id-theft-fraud-statistics-1276.php).
However, these numbers would be considerably higher if it weren't for the use of data and machine learning in the attempts to stop these criminals. Being able to monitor spending patterns, where people are buying and the kind of items they are purchasing is essential to being able to see if they are being targeted. It allows the biggest banks to not only protect their customers, but also their own capital as each fraudulent transaction costs them money. It is this kind of work that has seen considerable drops in the number of card ID thefts in the UK, with a 14% fall between 2013-2014 alone.
We are also seeing considerable developments in the way that machine learning is protecting individuals across the world through learning trademarks from previous attacks. It is something that we are seeing with IBM, who are now utilizing their Watson cognitive computing system to help in cyber security. According to IBM, there is a huge amount of information about cyber security online, but 80% of it is unstructured and not accessible by regular data mining systems. This makes cognitive computing incredibly powerful as it can not only take the structured data, but also this huge amount of unstructured data to help build the most effective security systems that protect against as many threats as possible.
However, the biggest single element that could help in cyber security is ironically through breaking down walls, rather than building them to keep hackers out. This is not in the sense of allowing hackers to have easier access, but more in creating an open data platform between different security companies. It is something that is not widely used, but is likely to increase as the threat from attacks increase.
What these companies are currently finding is that, despite demanding increasingly large fees for their services, the ability to keep out hacks isn't always there. Target spent $1.6 million on a security system and still lost the credit card details of 40 million people to a single Russian 17 year old, for instance. Hackers are increasingly finding it easy to get through systems that are siloed, who don't share their data and operate by themselves in the hope of gaining competitive advantage. They can essentially take the same approach to different systems, because data on the attacks isn't shared.
Sharing this information and sharing data across those companies who may compete, but ultimately have the same goals, would be incredibly powerful. If a hacker tried to get in to any computer through the security systems and was found, this would instantly share the data with every single security system on the planet. It would make it considerably more difficult for hackers and ultimately make everybody's systems more secure.