The increasing complexity and volume of cyber threats targeting mobile platforms, alongside the necessity for cross-platform adaptability between Android and iOS, underscore the urgent need for advanced artificial intelligence (AI), generative AI-based threat detection and malware detection systems. This research addresses the critical challenges of accurate threat detection, adaptability, and ethical considerations within mobile security frameworks. We explore the development of models aimed at general mobile threat detection, focusing specifically on malware and benign detection. By leveraging AI and generative AI, we delve into the complexities associated with these detection processes, emphasizing the importance of explainability in fostering user trust and compliance. Our findings evaluate the use of AI solutions to enhance mobile security while addressing the ethical implications of their deployment. This paper contributes to future advancements of mobile security by deducing avenues for further research, comparing Random Forest and XGBOOST models, ultimately aiming to create more robust and adaptable security systems that can effectively respond to the evolving threat landscape.