AI will make humans redundant in security, but that’s a good thing
The (human) panel
If security is too big a question for us to answer, maybe it’s time to hand the responsibility over to the machines, reports Telecoms.com (Banking Technology‘s sister publication).
During a panel at “AI and Machine Learning World”, part of tech event TechXLR8 in London, artificial intelligence (AI) was discussed. What might seem quite a controversial topic, is fast becoming a reality. AI might be the solution to the security conundrum. Since the inception of the internet and the invention of cybercrime, we’ve constantly been asking ourselves whether it is possible to remain secure.
“Life is becoming more digital, so I don’t see any reason why the crime in real life will not be transferred to the online world,” says Dr Bertrand Hassani, head of research and innovation for risk methodology at Banco Santander.
We’re always being told to be wary in real life, though this lesson seems to be forgotten in the online world, where we are perhaps at greater risk. If you lose your wallet in real life, you lose your cards and a couple of quid. If someone penetrates your online self, the repercussions could be disastrous. This is the same for both individuals and organisations.
“Education is a massive area,” says Patrice Slupowski, VP of digital innovation at Orange. “The user is the weak link most of the time.”
It’s a simple statement to make, but more often than not correct. We are the weakness in the chain and the root cause of an unhealthy proportion of data breaches or malware infiltrations. The question still remains however; is it time to hand the security reigns over to the machines?
There are two advantages here; firstly, computers are not prone to human error. An AI application won’t leave a laptop in a pub after a couple of beers. But, secondly, and perhaps more importantly, the computers can deliver more productively than we can ever hope to achieve.
The number of hackers are growing quickly. Whether they are out for personal financial gain, or are political activists hoping to expose some sort of scandal, the objectives are the same. Search the perimeter attempting to find some way to penetrate defences. With the number of software updates and the trend of mobility, as well as the recurring problem of shadow IT, there is bound to be a chink in the armour somewhere.
It would be unfair to assume security professionals are going to work around the clock to achieve the almost impossible task of finding the current frailties, while also predicting where the future ones will be. These guys are always reactive to breaches, they are always one step behind.
“Most attacks are opportunistic,” says Hassani. “Once we improve our systems and use AI naturally, we can start combatting them [i.e. the hackers]. We now stand a fighting chance.”
Using AI in security seems to be a natural progression. The machine can see patterns we can’t because of the volume of data which it can process, it can work tirelessly and predict where the system needs to be strengthened by accessing the vast information on the internet to learn about potential attacks. This is not a case of AI replacing humans, but achieving a task which humans simply couldn’t. It’s an impossible ask.
