Srikanth
The hype of Machine Learning is overwhelming in every possible way. Sure, Machine Learning can solve some very pressing issue of information security world, but it can not be the only cure. The cyber threats are growing every day. With a billion more devices that are ought to come online with IoT, it is going to be very difficult to contain he threats. Imagine one of your IoT smart device is hacked, since it is connected to your home network, all the devices in that network will indirectly gets hacked, unless and until they have pretty nasty security protocols. This is going to be even harder when 5G comes around, which is probably going to be next year. Machine Learning can greatly automate the process of threat detection and threat anticipation.
What does Machine Learning offer in solving cyber security problems of 2018?
With ML you can automate the threat detection to an extent. That gives the IT department to figure out the more challenging security problems, while routine threats can be handled by ML. If trained properly, ML can help out to lay a defensive security strategy for hackers lurking on the edge of the networks. Sometimes, it is necessary to lay out security strategies which can hurt the hackers. In that we not only protected our information but also caused some loss to others, so that they will not try again. This may not work always, but if you are planning to setup something like that, ML can be of great help.
Phishing attacks are one of the most underrated attacks in the cyber security world. We may set up firewalls that can ward of best lot of the hackers, but that one employee what can’t resist an urge to 50% discount on clothes can cause a lot of damage. ML, if trained to identify such content which is dubious can help you to build your security even more.
Predictions is one of the strong suite of ML, and this can be used in our advantage. Even before we begin to create security rules for our networks and devices, primarily we have to sort out a way to write algorithms to predict the vulnerabilities in our networks. Then we have to make sure we get regular updates on the vulnerabilities. This is one way to protect our information.
We can’t depend on Machine Learning extensively. It can only do so much. However, we can always use Machine Learning to our advantage. Try not to automate everything with ML, because even ML can give you bad predictions on threats and you will be the one to lose. Have some manpower sit at the end of the day to evaluate the progress. We can never depreciate the infosec professionals relying solely on ML interfaces.
