In this era of digital revolution, the size of networks with increased connected devices has inflated exponentially, so the risk of their exposure to security breaching attacks. Intrusion Detection Systems are potent to imbibe cybersecurity to the continuously expanding network. It monitors the network traffic and hosts activities of the networking system to which it is installed for detection of intrusions. With the growth and complexity of cyber threats, traditional IDS methods struggle to provide effective real-time threat detection. This paper explores the advancements in IDS, focusing on Artificial Intelligence based machine learning approaches to detect the intrusions with effective accuracy and minimized false positives. This study analyses existing IDS frameworks, evaluate their effectiveness, and propose an optimized model integrating anomaly-based techniques. For experimentation, four machine learning algorithms are deployed over KDD’99 IDS dataset. Data preprocessing is done with scaling and selecting the features. The proposed model is assessed for its performance over accuracy and f-measure. Experimental results demonstrate the proposed model’s efficacy for detecting a pretty good range of cyber intrusions. This study provides valuable insights into the evolving landscape of IDS and highlights future research directions for improving cybersecurity resilience.