Explainable AI for Android Malware Detection: Enhancing Transparency, Privacy, and Accountability in Cyberlaw

The rapid growth of Android applications has increased the risk of malware that threatens user privacy and digital security. Machine learning techniques have shown strong capability in detecting malicious applications, yet many existing approaches lack transparency, which limits their applicability in legal and regulatory contexts. This study proposes an explainable artificial intelligence-based framework for Android malware detection that integrates data preprocessing, feature selection, and a Random Forest classifier with SHAP for model interpretability. A dataset containing 4,465 applications with 242 features was analyzed, and extensive preprocessing was performed to remove duplicate and inconsistent samples, resulting in a refined dataset of 658 instances. The proposed model achieved an accuracy of 0.8712 and an AUC of 0.9274, with stable cross validation results indicating good generalization performance. The explainability analysis revealed that permissions related to sensitive user data, such as phone information access, SMS handling, and system control, play a significant role in malware classification. These findings highlight the importance of combining accurate detection with interpretability to support transparency, accountability, and compliance in cyberlaw. The study demonstrates that explainable AI can enhance trust in automated cybersecurity systems while providing meaningful insights for legal and regulatory evaluation.