5G networks are better than ever and, in doing so, upend connectivity with lightning-fast speeds and increased capacity, but thetechnology also brings with it a whole new set of threat vectors, including cyberattacks and data breaches, that need to be defended against. Many of the currently available intrusion detection systems (IDS) fail to achieve this balance, resulting in gaps in the detection and mitigation of in-line attacks asthey occur. The research investigates these issues by exploring data collected on the CIC-IDS-2017 dataset, as well as comparing the performance of a variety of machine learning and deeplearning models (BiLSTM, CNN, Random Forest, and hybrid approaches). Although CNNs show some promise in achieving a high accuracy for intrusion detection,it is the Random Forest classifier that shows considerable potential with an accuracy of 93.5%. Surprisingly, a hybrid model outperforms allothers, recording an accuracy of 95.02%.