Come, Let’s take a walk down the Dark Side !

Is it alright to say, that the last 5 years weren’t a great year for Cyber-Security ?

A lot of high end attacks like ransom ware were deployed on Tech giants like – Equifax, Uber, Deloitte has sent a chill down the spine of every CTO and highlights the responsibility of protecting the company’s most invaluable asset – Data. Despite the plethora of Security Companies that continue patching widely used software’s, attacks continue to rise.

Image result for growth of cyber attacks in 2018

With the rise of these attacks, It brings us to the question , what will the next 5 years of this industry look like ? What kind of threats would be battling ?


The Dark side of A.I.

Webroot, a Network security company based in Colorado, says, it used Machine Learning in approximately 90% around all of its products. The downside to this double edged sword is they also predict about 90% of security professionals have to deal with more sophisticated Cyber attacks powered by A.I.

For example – Machine Learning basically automates data collection, which is usually made or sourced from Open Software Platforms. If the hackers use AI to crack passwords for these technologies by limiting the number of combinations based on Location, Demographics etc; companies are playing a huge uphill battle.

Source – MIT, Hacker news
Analyzing Market Trends :
Some of the most recent trends in AI based Cyber security is continued adoption on of Cloud in Cyber Security and the use of Chatbots to counter Cyber-crimes. Cloud Security is booming as it seems to be the most “Cost effective solution” for Medium to large businesses as it removes the need to buy, and hire people to maintain these networks. An interesting area where a lot of investments are pouring in, include – AI capabilities for Cognitive Computing as it provides new gen Cloud Computing technologies the intelligence to combat cyber attacks. To give you a brief overview as to what these software’s do – Speech Recognition, Text Analytics and Sentiment Analysis to understand the intent of new code trying to breach into a network.


Chat-bots are surprisingly gaining a lot of traction in major companies, again , with a major focus on Cognitive computing. Chat-bots are trained for responses for a specific set of questions asked by the user.  Powered by dynamic translation & text analytics chat-bots are allowing companies to counter cyber-threats.
A periodic table view of the Landscape in Cyber Security, Source : CBINSIGHTS
I’d like to conclude this with a quote, something to ponder upon.
“We are giving away too much bio-metric data. If a bad guy wants your bio-metric data, remember this: he doesn’t need your actual fingerprint, just the data that represents your fingerprint. That will be unique, one of a kind.”  
— Mike Muscatel, Sr. Information Security Manager, Snyder’s-Lance, at SecureWorld Boston

2 comments on “Come, Let’s take a walk down the Dark Side !”

  1. That actually leaves a lot of room for improvement and development for cyber-security industry! I love your part about bio-metric data and I do think people don’t realize the importance of their personal data being stored, used or tempered with in various ways that is detrimental to their privacy security, and the more we emerge into the digital experience in life, the more privacy we sacrifice in exchange for convenience. I know this privacy issue is controversial, but I am not taking sides on this because there are advantages and disadvantages on both sides concerning privacy and comfort. It’s really interesting to dig into it/

  2. I do agree there is huge potential for AI in cyber security field. A good example will be using machine learning to detect abnormal activities from a device or an account. And also, it’s a good point that data nowadays collected by some companies actually gives away our privacy. There are tons of footprints we left on internet that others can use to analyze our behavior patterns.

    However, I am confused how AI can be used to crack passwords. The data dimension of passwords may be too high for ML algorithm to learn and it will need huge amount of data to train the model.


Comments are closed.