Artificial intelligence (AI) is reshaping the operation and management of energy pipelines that transport oil, natural gas, hydrogen, and carbon dioxide. This review provides a critical synthesis of current developments, highlighting how AI is being applied across pipeline life cycle stages—from route design and construction quality assurance to operation, maintenance, and eventual decommissioning. Recent advances in deep learning, transformer architectures, graph neural networks, and physics-informed models demonstrate the ability to detect anomalies, optimize multiphase flows, and support predictive maintenance with greater adaptability and robustness than conventional methods. Case studies illustrate how industrial operators are deploying AI platforms to improve efficiency, reduce environmental impact, and enhance system resilience. Alongside these opportunities, the review identifies systemic risks, including data bias, algorithmic opacity, cybersecurity vulnerabilities, and overreliance on automation, and discusses emerging mitigation strategies such as synthetic data generation, explainable AI, defense-in-depth frameworks, and mixed-initiative human-machine collaboration. Looking ahead, the field is moving toward multi-source heterogeneous data fusion, few-shot and transfer learning for data-scarce environments, and causal inference for decision support. AI applications in hydrogen and CO₂ pipelines, though nascent, offer promising directions for building safer and more sustainable infrastructures. This review underscores the need for continued interdisciplinary efforts to balance innovation with reliability, ensuring that AI becomes a trusted enabler of next-generation pipeline systems in the global energy transition.