<p>Generative artificial intelligence (GAI) advances are attracting wide attention as they continue demonstrating unprecedented potential for transforming civil and environmental engineering (CEE). Despite this potential, a limited amount of work examines the current state-of-the-art of GAI in our domain. In order to bridge this knowledge gap, we present findings from a scientometric literature review to identify the most commonly used GAI algorithms/architectures/models (including large language models (LLMs)), best practices, and barriers to adoption within CEE. Our review indicates that GAI's notable applications span complex tasks (such as automated structural layout generation, project planning, context scheduling, etc.) and iterative/routine assignments. In addition, our review pinpoints several challenges, including domain-specific data availability and heterogeneity, integration with existing engineering workflows, lack of proper datasets for model training, the need for systematic evaluation metrics, and codal provisions. We conclude by promising directions for future development.</p>