This paper presents an AI-driven approach for real-time detection of humans using an unmanned aerial vehicle (UAV). Leveraging the advanced capabilities of YOLOv8 object detection model implemented with the PyTorch framework on the Nvidia Jetson AGX Orin device, our methodology aims to enhance situational awareness and decision-making in critical missions. The implementation begins with the integration of the YOLOv8 model into the UAV system, enabling rapid and accurate detection of humans, allies, and enemies in dynamic environments. The Nvidia Jetson AGX Orin device provides the necessary computational power and efficiency to execute the YOLOv8 model seamlessly onboard the UAV, ensuring real-time updates and reliable target identification. Furthermore, we discuss the technical specifications and advantages of using PyTorch with YOLOv8 on the Jetson AGX Orin device, highlighting the ease of development, deployment, and scalability for artificial intelligence (AI) based applications in aerial surveillance and reconnaissance. Overall, our work contributes to advancing UAV-based target detection capabilities, enabling efficient detection of humans, allies, and enemies in real-time scenarios, thereby enhancing mission success and operational effectiveness in dynamic environments.