Artificial is a constantly evolving and progressing technology. In today’s society, artificial intelligence models are frequently employed. With today’s data and the capabilities of artificial intelligence, we can attain human intellect and, in some situations, considerably more. But also, the rapid increase of data and how it is managed is a cause for concern. Machine learning is an artificial intelligence sub-area that allows a machine to learn from prior data. Nowadays, machine learning is utilized in a broad range of applications. However, it is a tough task to apply machine learning on decentralized data. In the recent years, a new technique known as Federated Learning has garnered considerable attention. It concentrates to train distributed and decentralized data while maintaining its privacy. In our paper we have explained what Federated Learning is, the characteristics of Federated Learning, what are the hardships we face in applying Federated Learning, in which field Federated Learning is currently used and what are the future scope of Federated Learning.