As cities transform into smart cities, they are increasingly filled with Internet of Things (IoT) devices that bring new challenges to keeping networks secure. This paper explores how Artificial Intelligence (AI) can be a game changer for network security in these smart city environments. It discusses how AI can improve traditional security by providing real-time threat detection, automated reactions, and forward-looking threat analysis. We particularly look at how machine learning can power Intrusion Detection Systems (IDS) to spot unusual patterns in network traffic, helping to predict and mitigate potential threats more accurately. We also explore how reinforcement learning can dynamically tweak network settings to enhance security while efficiently using resources. These AI-driven techniques speed up response times compared to manual methods and boost the precision of detecting real threats while minimizing false alarms. This study highlights AI's vital role in safeguarding critical urban infrastructure like energy grids, transport systems, and healthcare networks. It also considers the complexities AI introduces, such as issues with privacy, potential biases, and the need for clear system transparency, pointing out that these issues require thoughtful consideration as we apply AI in smart city security.