AI Based Threat Detection System For Military Purposes with Integrated AI ChatBot

This paper presents an AI-based threat detection system designed to support military decision-making through intelligent image analysis and interactive chatbot assis-tance. By integrating YOLO-based object detection with a hybrid LLM-driven chatbot framework, the system addresses key limitations of conventional manual threat assessment methods, including delayed analysis, frag-mented reporting, and limited real-time decision support. Key contributions include: (1) automated threat detection using a fine-tuned YOLO model for military-relevant ob-jects, (2) structured threat scoring and risk classification into prioritized levels, (3) an integrated hybrid chatbot combining rule-based control with LLM-driven contex-tual reasoning, (4) secure role-based access architecture for controlled system interaction, and (5) near real-time processing with explainable threat summaries. Experimental validation demonstrates high detection per-formance (mAP = 0.91) and strong agreement (r = 0.82) between chatbot-supported recommendations and expert evaluations. The system achieves efficient end-to-end pro-cessing (2.8–3.4 seconds per image) while maintaining reliability and operational clarity. User feedback further confirms the usefulness of chatbot-generated summaries in assisting threat interpretation and response planning. Future work will focus on multimodal data fusion (image + text intelligence), edge-device deployment optimiza-tion, enhanced explainable AI mechanisms, and adaptive learning frameworks to handle evolving threat patterns. The proposed system represents a significant step toward intelligent, scalable, and decision-centric AI deployment in military threat analysis environments.

[1]     Shanmukh Shekar K C , Suresh Kumar Natarajan, Ambika I ” A Deep Learning Approach for Automated Image Captioning with CNN and LSTM Models ” In 2025 [2]     A.Meena et al., AI-Powered Threat Detection and Mitigation System: A Literature Survey [3]     Chhaya Dhavale ,Trupti Pawar , Aradhana Singh,Shubham Pole , Krishna Sabat Revolutionizing Farming: GAN-Enhanced Imag-ing, CNN Disease Detection, and LLM Farmer Assistant ” In IEEE Conference, 2024 [4]     Rahul Chauhan, Gurnain Singh Wadhwa,Deepak Kumar Verma,Krunal Vaghela,Chandradeep Bhatt Exploring the Intersection of Language Processing and Image Generation in a Conversational AI Application ” [5]     Devang Kumar Bharti , Shikhar Shukla, Shikhar Bhandari, Apeksha Koul. Conversational Interface for Textual and Visual Data Interpretation using customized CNN. ” [6]     Shengyang Su Conversational and Image Recognition Chatbot ”