Effects of artificial intelligence (AI) in software security

Abstract

The combination of Artificial Intelligence (AI) with software security has completely transformed the traditional cybersecurity practices and methods by offering advanced capabilities in different domains like threat detection, vulnerability prediction and automated defence mechanisms. This research aims to identify, describe and categorise the different factors of AI in a software security environment. This Systematic Mapping Study (SMS) is based on the research work from thirty-seven selected studies published between 2008 and sept. 2025. Through a very disciplined approach and rigorous methodology which involves different database searches, thematic coding and analysis of security challenges and their solutions, different tools used and ethical concerns that arise because of using AI. We have identified different key themes that use of AI in applications in cybersecurity environment for threat detection and prevention including the security of AI systems themselves and the role of generative AI. Different findings reveal and present data that AI enhances detection accuracy up to some certain percentage in spam and anomaly detection but also introduces challenges like insecure code generation, algorithmic bias and overreliance on automation. Ethical concerns including privacy risks and transparency issues are prevalent in 70% of studies. The findings of this study will help researchers and practitioners of this field to do more research studies on the integration of AI in the software security environment. This will also help in developing specialised solutions for AI applications in the software security environment.