Maintaining a secure company network can be a daunting task, and that’s putting it lightly. The number of cyber threats out there seem to be multiplying by the day, while the incidents of cyber attacks have become a common headline. Just ask Target or Sony
about the damage that can come from infiltration by cyber criminals and hackers. And those headlines appear only when a major company is the victim.
Now, imagine how many attacks don’t make the news. In 2015, the number of data breaches
rose by 8 percent when compared to 2014. The result was more than 700 million data records lost or stolen. And all signs point to this trend continuing.
Needless to say, enterprises and other organizations have their work cut out for them when it comes to securing their most sensitive information. But there may be hope. The best chances of success in improving cyber security could come from the combination of artificial intelligence and human security analysts.
For many businesses, security was best summarized as one of two options. First, they could adopt the latest security tools and allow those platforms to go to work. Advances in these technologies, particularly in the field of big data, have made this a distinct possibility that many companies are eager to try. The second option was usually hiring the best security experts to monitor their networks and devices, finding any incoming threats and eliminating them before they grew out of control.
While both of these approaches are the most common, it’s easy to see the flaws in each one. Relying solely upon security tools, even those that include machine learning techniques, can only be so effective. One of the main problems of only using machine learning is the tendency for the platform to indicate a high number of false positives. This can slow down operations and render security efforts far less effective.
It’s just one of the many challenges with big data
that companies have to face. Humans, on the others hand, won’t come up with so many false positives, but they are limited in their ability to identify the latest hacker techniques. They rely on rules set forth by the experts, making it more difficult to find threats that fall outside those rules. In a sense, they’re always playing catchup.
The key, it turns out, could be in combining the two. A new hybrid system
developed by researchers at MIT uses artificial intelligence techniques with human analysts to better detect cyber threats. While machine learning on its own only detects up to 73 percent of attacks, the new system is able to detect 85 percent of attacks while also cutting down on the number of false positives significantly.
The system works by going through lots of data and detects any suspicious activity
with an unsupervised predictive model. These events are then passed on to a human analyst, who can confirm if they are real attacks or not. The system takes this feedback into account and incorporates it in order to form a new supervised model. This continues, with each iteration improving upon the model.
Of course, no system is perfect, and the 85 percent detection rate can still be improved upon. Next Gen AI systems
also combine AI with human experts, helping the two work side-by-side for a common goal. The systems combine the fast response needed to thwart cyber attacks with the steady guiding hand of someone who knows if those attacks are genuine threats.
One aspect of AI and humans in action comes in the form of video analytics
, which is most common with security surveillance. A camera with intelligent video analytics can detect problems and recognize possible threats on its own but no actions are made on the system’s own. Instead, alerts are sent to security teams so they can check out the threat and determine if actions are necessary. In this case, detections are more effective and true threats are dealt with before they become a major problem.
The combination of AI and human skills may be only the beginning of a new approach to cyber security. Not only is it more effective, but it saves on the amount of work needed. The bigger picture also highlights that AI likely won’t completely replace human workers, as many have feared. AI and human minds are simply better when they work together.
In any case, organizations should have plans in place to minimize the damage from attacks and to get operations back up and running quickly. With this progress, security doesn’t have to seem like such a formidable challenge. Help is on the way, and just in time, too.
About the Author: Rick Delgado is a freelancer tech writer and commentator. He enjoys writing about new technologies and trends, and how they can help us. Rick occasionally writes for several tech companies and industry publications.
Editor’s Note: The opinions expressed in this and other guest author articles are solely those of the contributor, and do not necessarily reflect those of Tripwire, Inc.