The advancement in science and technology has become a significant bottleneck to the information/data. In order to overcome the threats, a number of information hiding methodologies have been proposed. Among such techniques, the most commonly used technique is information hiding in an image due to the presence of secret information in an image, where it will be hard to be identified by the attacker. The appearance of both the normal image and the image that contains the secret information will be similar. Also, with a number of deep learning and AI techniques, data hiding can be done in an increasingly better way. This paper summarizes various neural network-based image steganography techniques and also the methodology used in each paper. Also, this research paper discusses the advantages and drawbacks faced by each technique.