In 2014, Goodfellow et al. introduced Generative Adversarial Networks (GANs), an adversarial learning framework designed to generate synthetic data. In rolling bearing fault diagnosis and prognosis, specific GAN variants such as Conditional GANs (cGANs), Wasserstein GANs (WGANs), and their derivatives have been developed to address data scarcity and class imbalance challenges. In this work, we conduct a literature review to systematically examine GAN-based data augmentation techniques for fault diagnosis and prognosis of rolling bearings. Through a rigorous selection process, we identified 229 primary studies that employed GAN-based data augmentation, underscoring the widespread use of GANs to generate synthetic data in this field. Our review shows that GANs were first applied to rolling bearing fault diagnosis in 2018, and their use has grown significantly since then. Among GAN variants, Wasserstein GANs (WGANs) and Conditional GANs (cGANs) have proven highly effective in generating realistic synthetic data, particularly when integrated with Convolutional Neural Networks (CNNs). The review further reveals that CNN models have been widely used, achieving accuracy rates exceeding 95% in fault diagnosis and prognosis. We also report that 90% of studies employ accuracy as the primary evaluation metric, while 15% use F1-score, as detailed in our metric analysis for bearing fault diagnosis. For fault prognosis, RMSE and MAE are the most commonly used metrics, appearing in 11% and 9% of studies, respectively. Our analysis reveals standardized hyperparameter configurations with learning rate 0.0001, Adam optimizer, and batch size 32 being most effective. The review identifies critical challenges including data imbalance (19.7%), training instability (11.0%), and data scarcity (10.7%) as primary bottlenecks for industrial adoption. This review establishes a comprehensive foundation for understanding the current state and future directions of GAN-based approaches for rolling bearing fault diagnosis and prognosis, offering researchers and practitioners a valuable resource in industrial predictive maintenance.