PREDICTIVE ARTIFICIAL INTELLIGENCE MODELS IN INFORMATION SECURITY: CONCEPTUAL SYNTHESIS, CHALLENGES AND FUTURE TRENDS (2020–2025)

síntese conceitual, desafios e tendências futuras (2020–2025)

Abstract

The increasing complexity and sophistication of cyber attacks highlights the need for proactive and predictive approaches in information security. This article presents a conceptual synthesis on the application of predictive Artificial Intelligence (AI) models from 2020 to 2025, focusing on usage categories, key challenges, and emerging trends. The analysis is based on a narrative literature review of specialized sources, including IEEE Xplore, ACM Digital Library, Springer, and ScienceDirect. Predictive cybersecurity models can be classified into five main groups: supervised models, deep neural networks, federated models, hybrid models, and specialized models. Critical challenges include model privacy and explainability, integration with legacy systems, and data limitations. Emerging trends point to lightweight models for edge computing, autonomous multi-agent systems, and the implementation of explainable AI, aiming to enhance reliability and decision-making transparency. The proposed conceptual framework organizes these technologies and connects them to protection strategies, contributing to the advancement of knowledge in predictive cybersecurity. This approach underscores the relevance of actionable practices in corporate environments and provides a solid foundation for future academic research and practical applications in information security.