The Malicious Use of Artificial Intelligence in Cybersecurity

| Comments

Cybersecurity is a crucial consideration for any business or organization operating today. Those who are somewhat familiar with the cybersecurity industry will know that, until recently, there was a generally very lax attitude towards the subject, sometimes alarmingly so.

Naturally, as our cybersecurity capabilities increased, cybercriminals have begun turning to increasingly sophisticated methods to carry out their attacks. There are two components to any cybersecurity system. There is the system itself, the software and computers which comprise it and run it, and then there are people.

Traditionally, people have represented the weak points of these systems. Even though we can design computerized security systems which are nearly impossible to breach, all it takes is one person being careless with a password to bring the whole thing down.

But computers are becoming more powerful and more capable. Artificial intelligence and machine learning which we have used so successfully to keep ourselves protected from cyber attacks are now being adopted and exploited by cyber crooks to thwart our efforts.

Defining AI

Most people have heard of artificial intelligence (AI), and could probably list at least a couple of its existing applications. Those of us who grew up pre-2000 will remember when AI was new and exciting when we had high hopes and somewhat unrealistic expectations. In today’s world, where we are so used to many of AI’s most common applications, we don’t always think of them as being examples of AI.

Current applications of AI range from Amazon and Netflix recommending products and services on the basis of your previously stated preferences, all the way up to military prototypes for autonomous combat systems. Of course, most of the uses of AI fall in between these two extremes when it comes to their real-world applications. In the realm of cybersecurity, AI is used both offensively and defensively.

Good Versus Evil

One of the key considerations, when we are discussing the role of artificial intelligence in cybersecurity, is that AI itself has no moral bias. This means that an AI can be designed to help other computer systems and people, or it can be set to work against them. The technology itself is ethically neutral, leaving it to the architects and users of these systems to imbue them with an ethical slant.

As our understanding of AI deepens, and the hardware available to run the technology on also continues to grow in power, both the defenders and the attackers of computer systems have more to work with.

The biggest concern, from the perspective of cybersecurity professionals working to counter malicious actors, is that an increasingly sophisticated understanding of machine learning (AI that can learn how to learn) means that our most advanced defenses are rapidly rendered obsolete. Not only this, but the rise of machine learning (ML) could also lead to a sharp increase of zero-day attacks, attacks which remain unreported to the cybersecurity community.

Malicious AI

There are a few ways that ML is used to undermine our best security systems. For example, a sufficiently advanced ML-algorithm could, in theory, analyze the algorithms that a piece of antivirus software uses when determining if a code is malicious or not and find ways to circumvent them. Such an attack would be very sophisticated and depend upon access to privileged information regarding the cybersecurity software.

Another way that AI can undermine antivirus software is by poisoning data. We mentioned earlier that many antivirus systems rely on a regularly updated list of known malware and viruses. By inserting false positives into these databases, and selectively deleting other entries, malicious third-parties can hijack the learning process and prevent the software from successfully identifying future malware.

Staying Safe

In addition to the usual antivirus software, if you want to keep yourself secure, you should invest in a VPN service. A VPN will ensure that your online traffic is encrypted at all times, making it hard to trace and attack you. Many of the best VPN services also come with a range of other tools and features for enhancing your security and protecting you from malware.

In the future, we can expect to see AI and ML playing a much more prominent role in both defending our systems and attacking them. Most technologies can quickly be put to nefarious use. It is always essential to be mindful of the nature of the Pandora’s boxes we are opening.

Author Bio: Harold is a cybersecurity consultant and a freelance blogger. His passion for virtual security extends way back to his early teens when he aided his local public library in setting up their anti-virus software. Currently, Harold’s working on cybersecurity campaign to raise awareness regarding virtual threats that businesses have to face on a daily basis.