Breast cancer is the most common cancer among women. It is a disease in which abnormal breast cells multiply uncontrollably, forming tumors. These tumors can spread throughout the body, leading to serious and potentially fatal complications. Early detection of breast cancer makes treatment more effective and increases the chances of recovery. Artificial intelligence plays a crucial role in disease diagnosis, particularly in breast cancer detection. The data collected can be used by machine learning to create predictive models for breast cancer. However, not all of the collected data is relevant, which brings us to the concept of dimensionality reduction. In this work, we propose a technique based on autoencoders to reduce the dimensionality of the feature space, thereby improving the performance of machine learning classifiers. We train an autoencoder and use only the “encoder” part, which outputs the latent space where the most relevant features are concentrated. This latent space then serves as input for a set of machine learning classifiers. Our results are promising, as our approach improves the performance of machine learning classifiers while reducing their training time.