Deployment of large-scale grid-connected PV systems continue to face issues around and challenges of reliability, stability, and quality of power because of the intermittency, and nonlinearity of solar resources. Intermittency and nonlinearity of solar resources also make the implementation of strategies based on fixed-rule and model toolsets increasingly inadequate. This paper documents the benefits, and the results of employing intelligent, Artificial Intelligence (AI) technologies for analysing and building toward increased reliability and stability of interconnected PV systems. The investigation examines the operative current and post AI and ML (machine learning), deep learning, and reinforcement learning frameworks built around system functionalities of fault monitoring, diagnosis and control, dynamic inverter control, power forecast and control, and voltage and frequency control, as well as grid support and self-sustaining systems. Proactive and reactive AI driven frameworks and machine learning methods from predictive control have been tested on systems to develop fault detection, and dynamic control under operational challenges. The challenges of control and uncertainty dynamically and operationally based on real world datasets and simulations. numerous active disturbances, voltage dips, frequency decline, integrated control of partial cloud systems and rapid change of intelligent solar system Shadow loss of energy. The achieved intelligent systems brought numerous challenges on operational latency. Cyber control and information system security challenges have been eased to provided operational state. The intelligent systems have sustained the operational uncertainty expected. The paper provides and elaborates on the numerous challenges attained. The findings enable utility companies and policymakers to view and appreciate the role of AI as an essential component to sustained and adaptable grid-tied solar PV integrations