Artificial Intelligence and Machine Learning Are the Future of Cybersecurity

According to a report by Accenture Security1, ransomware, alternative crypto-currencies and highly-evolved deception tactics are the top cyberthreats of 2017. The internet is also overrun by bots2 — these automated programmes, many of which are malicious, contribute to more than half of the world’s web traffic.

Hackers are continuously innovating new ways to gain access to networks as well as steal critical data and information. With cyberthreats rising exponentially, each becoming more sophisticated than the last one, how can companies fend them off, and quickly enough so as not to suffer interruptions to their operations? Not only that, the cybersecurity workforce shortage is a burgeoning problem, set to reach 1.8 million in talent gap by 20223. Manpower and expertise in this field will be hard to come by.

Cybersecurity automation: the new frontier  

The panacea to these woes is automation. Today’s cybersecurity businesses are relying on artificial intelligence (AI) and machine learning (ML) to fortify cyber defences. AI is a type of computer software programmed to work and react like humans. It can easily and speedily handle an endless volume of data, recognising patterns and even the smallest anomalies. When put to work, it helps to heighten vigilance and reduce human errors. Cybercriminals could gain access to a company’s network via a phishing scam, and analysts may dismiss this breach as minor. Yet this breach could be the start of a domino effect, eventually triggering a full-on cyberattack.

As impressive as the technology sounds, AI is not entirely foolproof. Hackers have already figured out how to circumvent AI. Moreover, we are still light years away from building the perfect AI model, one with the thinking capacity of a human brain. That is why AI is often used in tandem with ML. As its name suggests, ML models can learn and adapt on their own. Tapping into data that represents samples of behaviours, ML algorithms build models that ape the behaviours of real-world systems4.

By learning about previous incidents, these models can then predict outcomes, understand traffic patterns and user activity, amongst other tasks. This is a boon as malware or phishing campaigns take on a different appearance or form with every attack. Well-trained models can sift through and classify such new, unlabelled data.

While much traction still has to be made before AI and ML can promise guaranteed success, the application of the two in cybersecurity is already paying off. A global security study by Nemertes revealed that it typically takes companies 10 minutes to an hour to detect a cyberattack5. With AI and ML, this time is whittled down to zero.

Collaboration the key to safeguarding against cyberthreats

Any business that has a digital footprint is subject to cyberattacks. During the SMU Cybersecurity Forum held in April 20176, Professor Robert Deng, School of Information Systems, Singapore Management University, implored government agencies, tech companies and private organisations to tackle the issue as one.

Both the public and private sectors must continue to invest in cybersecurity research, delving into areas such as cryptography, network security, data security and security management, the professor opined. Raising awareness through public education is also of paramount importance, as users must be able to spot and avoid phishing sites, malware or ransomware.


1 Accenture Security Report Identifies Top Cyber Threats of 2017:

2 The Internet Is Mostly Bots:

3 (ISC)² Cybersecurity Workforce Shortage Continues to Grow Worldwide, to 1.8 Million in Five Years:

4 Separating Fact from Fiction: The Role of Artificial Intelligence in Cybersecurity:

5 How AI can help you stay ahead of cybersecurity threats:

6 Staying ahead in the cybersecurity arms race: