7Explainable Artificial Intelligence for Cybersecurity

P. Sharon Femi1, K. Ashwini2*, A. Kala1 and V. Rajalakshmi1

1 Sri Venkateswara College of Engineering, Sriperumbudur, Tamil Nadu, India

2 Amrita School of Computing, Amrita Vishwa Vidyapeetham, Chennai, India


Artificial Intelligence incorporates human intelligence into machines and builds statistical models for providing solutions in various domains like cybersecurity. Though the AI-based techniques for detecting cyberattacks and threats are more efficient than traditional cybersecurity techniques, they lack explainability. These methods provide black-box solutions that are not transparent and interpretable in understanding the steps involved in reaching specific predictions. This reduces the confidence of users in the models used for cybersecurity, particularly in the present scenarios where cyberattacks are becoming increasingly diverse and complicated. To overcome this, Explainable Artificial Intelligence (XAI) is applied, which eventually replaces the traditional artificial, machine learning, and deep learning algorithms that operate as a black box. Given that cyberattacks are increasing day by day and the traditional AI algorithms are not sufficient for providing security, it is essential to focus more on XAI for exploiting the AI algorithms. This chapter provides a detailed discussion about explainable artificial intelligence and how it is applied in providing cybersecurity.

Keywords: Explainability, XAI, ...

Get Wireless Communication in Cyber Security now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.