The AI Arms Race: How machines are learning to fight alongside us in Cybersecurity


It can be useful to think of yourself as being under attack on your so-called digital territory all the

time. The opponent does not rest and regularly employs various kinds of attacks, from the complex

phishing to the brute force attacks which makes it really hard to defend oneself or some valuable

possession. This is the daily experience of cybersecurity specialists who are always aware of new

emerging cyber threats. But there’s good news: They assist in the delivery of the message of the

song which is no longer lonely. Thus, AI is emerging as a friend, not an enemy, as it is transforming

the landscape of cyber defence.


Consider AI as an extra-terrestrial security guard with extra-terrestrial level of strength and

concentration. It can process far more data than any analyst – network traffic (including encrypted

traffic analysis), user and system activity logs, even email communication schemes. Through this

approach, AI is not only able to detect anomalous behaviour but also analyse previous attacks,

including zero-day attacks, to make future predictions.

Multi-Layered Threat Detection: It can look at the data in different layers of a network security

system and determine there are relationships that suggest a multi-stage attack plan. It also assists in

identifying other vulnerabilities that a conventional signature-based detection system can often


Advanced Threat Hunting: AI can sort through an enormous amount of data to look for signs of

potential threats that may be concealed in a system. It can detect novelties in the usage of users,

network traffic, or perhaps in the use of system resources as a sign of an attack.

Deception Techniques: AI can be utilised in developing decoy systems or virtual honeynets in

order to ensnare attackers and gather information on their techniques and instruments. This

information can then be used to increase the overall cyberspace protection.


Beyond Malware Blocking: However, Check Point recently implemented an AI system which is

relatively impressive but remains just one option. Scientists are working on creating AI that matters

not only in stopping malware but also in identifying other types of attacks like ransomware attacks

and social engineering attacks.

The Rise of Adversarial AI: Now, hackers are also devising their AI tools to slip through the

defences put in place by the former. This includes malice oriented strategies such as adversarial

machine learning in which the attacker distorts data to obfuscate AI.

THE FUTURE: A POWERFUL PARTNERSHIP AI makes things easier yet it is not the be all and end all. The attackers’ tactics will also adapt as our AI defences become more robust and the AI arms race continues. This provides an endless cycle, where one side tries to outcompete the other one at their own game. The key to winning? Collaboration.

Continuous Improvement: Thus, it is imperative that AI security system development be a

continuous loop. Security researchers will be required to monitor the behaviour of an attacker and

refresh their models constantly.

Explainable AI for Security Teams: Since AI is increasingly implicated in security decisions, how

this is done (explainable AI) is integral. This will help in the decision making process of security

professionals and not mere reliance on the AI technologies.

Human Expertise in the Loop: AI performs well in the process of automation and decision-

making based on data handling, while the human factor is still crucial for risk assessment, decision-

making, and crisis management and forecasting. Security professionals will use AI for speed and

strength they possess, and they will step back and decide for other situations.


The guessing games that are still going on between cybersecurity organisations do have future

impacts. Therefore, we need to keep awareness of improvements, have a good partnership between

human and AI, and enhance our barriers consecutively to create more advancement in our digital

security. In unison, people will be able to come up with ways and means on addressing the issue of

cybercrime and hence we shall be able to enjoy a favourable security.

About the Author

Krishna Raj M

Krishnaraj is a cybersecurity enthusiast who began exploring hacking at the age of 16 and gained hands-on experience in penetration testing. He actively participates in capture the flag events, where he sharpens his skills with real-world challenges. His primary interest lies in red teaming, focusing on simulating sophisticated attacks to strengthen defenses and contribute to advancements in cybersecurity.

LinkedIn profile