Cyber security is a hot topic, to say the least. Incredible violations of privacy, government intrusions and numerous hacking scandals made cyber security a mainstream topic. We even have a TV show “Mr. Robot” about it. But do all these things make the general public more proficient at warding off the dangers of cybercrime? Not that much. Must I remind you that almost 20% of people are still using ‘123456’ as their password for various online accounts?
This already says a lot about the computer literacy of many people. That’s why cyber security is projected to grow up to $170 billion by 2020 . The expansion will continue until people actually learn how to handle their data or keep their eCommerce sites properly patched. And what does it all have to do with artificial intelligence? Everything. AI will shape our daily lives: from the way we do online shopping to the way we monitor our health. Cyber security is not an exception.
With 1 zettabyte (1 trillion gigabytes) of data transferred in 2016 alone, it’s impossible to build static algorithms that can automatically protect all of it. It has to be an adjusting, self-teaching and ever growing programming product that transcends an ‘anti-virus’ label and creates a new reality, where battles are fought with binary, not with artillery. Let’s take a closer look at AI’s role in cyber security.
Humans Are Slow
That is a fact. According to Moore’s law, computers will keep getting more powerful. Of course, we’re almost at the limit of how small of a transistor computer engineers can create, but there are certain achievements in the field of quantum computing, which might break this physical barrier.
We need to realize that computing becomes more and more efficient. Five years ago, nobody talked about artificial intelligence. And now, every other article or research from various domains of human interest deals with AI, automation and how all of these things are taking away jobs (with around 50% of American jobs being at risk) and reshaping economies.
And that’s all for one simple reason – humans are inefficient compared to machines. Sure, at this stage, this fact is debatable. But millions of jobs around the world are already replaced with AI or its basic forms. Cyber security is not an exception.
It’s a lot easier to employ a vigilante designed in line with the application security consulting standards, that will protect your data 24/7/365 without any need for food, accommodations or payments. This machine will never get tired and will always be at its peak performance, as long as your hardware is on par with its ‘appetites’. Not to mention that there’s a deficit of actual cyber security experts out there.
It’s a Two-Way Street
Cybercrime attracts immoral, yet brilliant minds. For many various reasons. Some just like the challenge, and some like ‘easy money’ (if they get away). But one thing is clear – they do understand the importance of AI in their operations. It’s their job to evolve.
People who are there to stop them have to come up with solutions that prevent cybercrime. Cyber criminals on the other hand have to come up with new attack methods that are able to circumvent the defense methods that are already in place. Imagine cybercrime as a bacteria and cyber security as penicillin. It’s up to the bacteria to develop immunity to that specific antibiotic. Otherwise it will not survive.
Given this paradigm and the fact that computing power becomes more available and robust every year, cyber criminals are definitely looking at ways AI can benefit their ‘craft’. Right now, it’s still mostly human vs. human. But this can soon change. With 30 security breaches per second happening in the world, cyber security experts can’t survive without AI. And cyber criminals won’t either, if this status quo isn’t challenged. Just look at our stock market – algorithms trading against other algorithms. This is what cyber security will look like pretty soon.
As we mentioned before, employing a machine is cheaper than employing a human. But it’s not only about the direct costs.
Think about the associated losses that cyber security is there to prevent. If the latest hacking scandals has taught us anything, it doesn’t matter how qualified or ‘protected’ someone is – there’s always room for an error.
Simply because we’re humans. All you have to do is to click a ‘bad’ link or download a ‘wrong’ file. And that’s it. You lose data, commercial information and, as we can see, even political leverage. That’s the cost of humans running the show, which isn’t even in their language – it’s either in a one (1) or a zero (0).
IoT is Like a Petri Dish
Think of the household devices that constantly get added to the Internet of Things. IoT is projected to triple and reach 30+ billion devices in 5 years. Think about the massive amounts of data that these devices collect for their main functions. Your appliances can communicate with the outside world, which makes them a standalone ecosystem within your household. Sure, that’s not true for many countries where progress is slower, but in a decade or so things will change dramatically.
All of these products will be perfect targets for cyber criminals. A hypothetical fridge connected to the Internet is not exactly a bastion of cyber security. But it can feed a lot of data to the perpetrators. The more products like that you own, the more data can be collected. It’s the perfect breeding ground for an invasion of privacy and cybercrime.
An AI system controlling these devices and working as a central hub, a neural center for your IoT belongings is the perfect solution for the upcoming onslaught of cybercrime, which will target specifically the IoT ecosystem. AI is a game-changer for IoT and its security.
Big Data is the Perfect Food for AI
Data science is a very complex niche, which serves as a great instrument for cyber security. It deals with large amounts of data, calculating risks and building predictions for possible vulnerabilities. That’s why banking and insurance industries actively use data science – it’s a reliable source of information for business decisions, including security.
The biggest problem is that data science on its own is limited by the people who apply it. That’s why at least 10% of commercial and government organizations in the US have already switched to a security system that incorporates elements of artificial intelligence.
Artificial intelligence is the perfect tool for enforcing cyber security rules. Machine learning, as part of AI-based solutions, is here to automate and simplify analysis, prediction and verification of cyber security risks. And the more big data you give it, the more efficient it becomes.
And that’s why it plays an important role in making the right predictions.
Prediction Is the Key
Predicting vulnerabilities is also a task that sometimes can only be accomplished with artificial intelligence. Especially if you have a very complex system that has a variety of possible ‘entry points’ and hidden vulnerabilities, which Donald Rumsfeld would refer to as ‘known unknowns’. Given the potential of machine learning to become more efficient with more incoming data, it’s absolutely vital for cyber security, especially with possible AI vs. AI scenarios that we discussed earlier.
Artificial intelligence is definitely going to rule the world of cyber security due to its superiority in multitasking, ability to increase precision through machine learning and increased efficiency due to vast capabilities that are only limited by the available hardware, which becomes better every year.