The advances in text generators that use AI today, such as ChatGPT, have made content dissemination and creation easier. We are now facing a wave of thousands of articles daily, although this facilitates access to information. On the other hand, it increases the risk of large-scale dissemination of false information. As a result, the problem of fake news becomes even more complex in the face of the massive volume of texts, as traditional fake news detection methods struggle to keep up with the exceptional growth of data. In this article, we address this issue by exploring the use of machine learning methods, more precisely the XGBoost algorithm and logistic regression combined with various vectorization methods, with the objective of training these two on a large dataset, observing their reactions, and trying to balance their parameters to achieve better results in handling large amounts of data. The objective is to equip detection systems with more robust tools to manage the abundance of content and strengthen defenses against disinformation.