This paper investigates the most prominent lines of optical network control evolution, focusing on software-defined networking (SDN), NETCONF/YANG protocols, telemetry techniques, advancements in packet/optical networking, and the integration of artificial intelligence (AI) within optical networks. We show how the integration of SDN with open modeling frameworks allows to devise hierarchical control models where we trade-off between the segregation of proprietary hardware and the creation of open interfaces like in the OpenSDK scenario. In addition, we depict the convergence of packet and optical layers with advancements in coherent technologies and pervasive telemetry techniques to create new flexible scenarios for controlling optical networks. On top of these approaches, the intent-based networking allows to implement configuration solutions using natural primitives. Finally, key applications of AI, mainly machine learning (ML), including quality-of-transmission estimation, failure prediction, and resource optimization, are analyzed to improve optical network control efficiency alongside their challenges, such as energy efficiency and data scarcity. By addressing advances in the aforementioned areas of research, this work outlines the transformative potential of combining programmability, real-time telemetry, and AI to build resilient, adaptive, and sustainable optical infrastructures for the future.