EXPLICATE: Enhancing Phishing Detection through Explainable AI and LLM-Powered Interpretability

AI transparency can match performance: EXPLICATE achieves 98.4% accuracy in detecting phishing attacks while providing clear explanations of its decisions, demonstrating that explainable AI doesn't require sacrificing detection effectiveness compared to traditional "black box" machine learning systems.

Users need to understand why AI makes security decisions: The paper emphasizes that lack of transparency in phishing detection systems "erodes the trust of users and diminishes their effective threat response," highlighting that security tools are only effective when people understand and trust their recommendations.

Plain language explanations matter: The framework uses an AI language model (DeepSeek v3) to translate technical explanations into accessible natural language, with 94.2% accuracy in explanation quality and 96.8% consistency with actual model predictions—showing that making AI decisions understandable to non-experts is both feasible and important.

Explainability can be practically deployed: EXPLICATE was built as both a usable desktop application and a Chrome browser extension, demonstrating that explainable AI systems aren't just theoretical concepts but can be integrated into real-world tools that people actually use.

Bridging the AI-trust gap is critical for security: The paper's core finding is that "high detection performance can go hand-in-hand with meaningful explainability," suggesting that responsible AI in cybersecurity requires closing the gap between automated decision-making and human understanding.