With the rise of artificial intelligence (AI), there is a growing prevalence of AI-generated images, making it increasingly difficult to distinguish them from real ones and this affects our society in many negative ways as well. This work develops a system that uses a Raspberry Pi with a Raspberry Pi camera and a machine learning (ML) model to identify AI-generated images in real time. Many models were tested using various feature extraction techniques such as Fast Fourier Transform, Discrete Cosine Transform, Wavelets and the best one is taken. It analyzes images captured by the Raspberry Pi camera and leverages the ML model to detect distinguishing features indicative of AI generation. The approach ensures quick and accurate detection under various lighting conditions and in real-world conditions, supporting the authenticity of digital content. Through successive work, an accuracy of 70% has been achieved in classifying AI-generated images, demonstrating the robustness of the model in different scenarios. It ensures the authenticity of digital content and highlights the feasibility of using low-cost, edge devices for this task. This work provides a foundation for advancing real-time image authentication systems in the era of generative AI. Future research can aim to improve accuracy through advanced feature engineering, larger datasets, and domain-specific optimizations as well as deal with the limitation of capturing the images through a camera, which might be misinterpreted as a real image by the model. Additionally, addressing evolving AI techniques and expanding the system’s applicability to videos and other multimedia content will further improve its utility and societal impact.