Fake news on social media platforms is increasing rapidly, so many people are becoming victims of this news without their interference. It is a big challenge for us to detect who is spreading fake news. Fake news spreads faster nowadays than in the past due to the widespread use of the internet. This research paper is a study of techniques based on artificial intelligence, such as neural networks, natural language processing, and machine learning algorithms that work together. The learning models surveyed are Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Bidirectional Recurrent Neural Network (RNN) methods. Natural language processing methods contain the tokenization model, and machine learning includes Term Frequency-Inverse Document Frequency (TFIDF) and unsupervised algorithms. The algorithms are compared and their effectiveness in detecting fake news is investigated, along with the advantages and disadvantages of the respective techniques.