SQL Injection Detection and Prevention Techniques using Traditional and AI-Based Approaches

SQL injection (SQLi) remains one of the most persistent and damaging threats to web applications, allowing attackers to manipulate database queries through vulnerable input fields. This survey paper investigates traditional and AIbased SQLi detection techniques, categorizing them into rulebased, machine learning (ML), deep learning (DL), and hybrid models. The strengths and weaknesses of each approach are reviewed using recent literature, with a focus on detection accuracy, precision, and false positive rates. A conceptual architecture is proposed for an AI-powered Web Application Firewall (AI-WAF), which integrates ML, DL, and Explainable AI (XAI) to enhance detection effectiveness and interpretability. Performance comparisons from existing systems such as ModSecurity, Snort, and DeepWAF demonstrate the potential benefits of intelligent, adaptive frameworks. The proposed AIWAF is not experimentally implemented, it is envisioned to assist future research in building scalable, real-time SQLi defense systems.