Generative AI is a type of artificial intelligence that, generates content based on input text (and sometimes images or sound). It processes the input carefully to discern valuable information from data that can be tossed away. The system takes care of this automatically, understanding with precision how certain details be recognized properly as well. Due to this, what the model outputs is correct but private with a content preserving serialization of the original input. A review of the most recent progresses in generative AI. This literature survey we cover advances, applications, and challenges to develop deep neural network as a powerful tool for generative modeling. In this paper, we review ten recent works and discuss the state-of-the-art techniques which can upgrade reinforcement learning models through a variety of perspectives to raising their efficiency, robustness, and extensiveness. Based on their wide usability in fields such as medical imaging, language translation gaming and even creative domains we can deduce the far-reaching consequences of generative AI. Still, there are substantial ethical and technical regulatory challenges to overcome for the technology to be used responsibly and effectively. This review highlights future research opportunities and stresses the importance of collaboration across different fields to overcome current limitations. The paper concludes by discussing the implications of generative AI for various industries and suggesting a structured approach for its continued development and ethical use.