Cybersecurity: Machine Learning at the Heart of Concerns in 2018
In a landscape marked by the explosion of the number of security incidents, Machine Learning should be at the heart of all concerns in 2018. The promises of Machine Learning are of interest to both hackers and companies concerned to protect their information heritage.
Machine Learning as a new battleground
Identified as one of the key issues of the year, Machine Learning finds very diverse applications in the world of cybersecurity. It is used, for example, to analyze all the activities performed with an authentication service so as to give the alert or block privileges as soon as abnormal behavior is detected. In this context, the system will study all the parameters of the connection attempt and seek to establish all the relevant correlations to decide whether or not to allow it. Here, it is the ability of systems to collect and process large volumes of data in real time that confers a form of intelligence to the machine.
On the other side, the attackers are not unaware of the benefits of this approach and exploit it on their own to test the existence of vulnerabilities or industrialize their social engineering campaigns. Their efforts come from new tools that can learn and adapt to exploit vulnerabilities more effectively. It remains to be seen which channels will borrow these attacks?
The other flagship trends of 2018
Marked by major offensives such as WannaCry or BadRabbit, 2017 saw an increase of more than 50% in the number of ransomware attacks. In 2018, McAfee estimates that hackers will tend to have fewer attacks, but better select their targets to maximize the chances of profit. The market would therefore shift from a volume-based logic to more sophisticated tools aimed at the most lucrative victims. Smartphones should be among the new favourite targets.
Special attention should also be paid to new applications distributed by one or more cloud providers in a “serverless” logic. This new way of consuming the resource on demand introduces new risks in terms of security: each new application used constitutes a new vector of potential attack.
The last part of these trends is the protection of individuals against the threats posed to them by the proliferation of personal data, favoured in particular by the democratization of the Internet of Things. We need to consider especially two aspects of the phenomenon:
- The first one is excesses, including marketing, that may arise from the exploitation of this information by the manufacturers of the devices concerned.
- As a corollary of the previous point, it also underlines the often poorly controlled crucial weight of consent granted by the end-user of online services involving personal data.
Machines and men
Result of the races? More than ever, cybersecurity will affect both machines and men in 2018. The former will have to learn to cope with increasingly sophisticated attack and defense techniques. The second will have the obligation to learn to control the use that is made of their information.