The issue of fake news has developed a lot quicker in the ongoing years. Online media has drastically changed its scope and effect all in all. On one hand, it's easy, and simple availability with a quick portion of data draws more consideration of individuals to peruse news from it. Then again, it empowers widespread of fake news, which are only false data to misdirect individuals. With 14 percent of individuals conceding that they have purposely shared a fake political report on the web, unmistakably these false reports will keep on picking up footing insofar as individuals are as yet ready to share them on the web. Subsequently, computerizing counterfeit news recognition has gotten urgent to keep up strong on the web and web-based media. There have been many machine learning approaches implemented to address and solve this problem. Likewise, this project, given a certain number of instances shared publicly on a social media platform, helps in tracing the instances one by one and identifies the fake news using supervised learning algorithms. It then generates news implying mixed viewpoints by applying adversarial machine learning algorithm and helps in creating a diversion. Thus, it reduces the impact created by misleading information. The prime focus of this paper is to compare the various existing Machine Learning, Deep Learning, Adversarial Machine Learning algorithms which will aid a researcher to understand the fake content detection spreading in the social media which are related to the educational domain. This survey paper will be useful for students during pandemic and crises like the covid-19 pandemic.