The integration of Artificial Intelligence (AI), the Internet of Things (IoT), and fifth-generation (5G) networks is fueling digital transformation within industries and societies. However, unprecedented connectivity increases the attack surface of modern communication infrastructures, putting 5G networks at risk of advanced cyberattacks and malware. 5G networks consist of dynamic, large-scale, and heterogeneous traffic, thus, the high dependency of traditional intrusion detection systems on signature baselines will not be effective. The present study addresses the issue of designing an SDAE-based framework for intrusion detection capable understanding and adapting to complex, high-dimensional traffic patterns and adapting traffic patterns within the 5G network. Most prior works concentrating on the IoT showed the 5G network's core capabilities that remain central, including network slicing and virtualization. Testing the proposed framework has produced positive results for 5G traffic simulated in the NSL-KDD dataset, achieving an average precision of 95.5%, a recall of 94.4%, and an F1-measure of 94.9%. These results point to the effectiveness of the framework in detecting and classifying anomalous activities in complex 5G network environments and show the framework's robustness and ability to generalize.