Machine learning has proven to provide important information to identify and predict areas with As and F safe groundwater. Statistical techniques and GIS were applied in the African Rift Valley to map fluorosis health risk zones and their relation to the geology; and to identify fluoride hotspots and coolspots in Tanzania (Ijumulana et al., 2020, 2021). A machine learning algorithm was used to recognize the main predictors of As rich groundwater zones in India (Mukherjeeet al., 2021). Toxic levels of As and F occur in aquifers at many zones of Mexico. This is a health concern since groundwater provides more than 60% of potable water. Areas with high As levels are related mainly with thermalism, geochemical processes releasing As in alluvial aquifers, often intensified with intensive abstraction, and presence of As bearing minerals in mining zones. These processes take place in diverse hydrogeologic and geological settings. Determining and predicting As and F groundwater rich zones through machine learning implies a challenge in a country with a complex geology, tectonics, and hydrogeological characteristics; active volcanoes and geothermal zones. Reliable and enough information of As and F concentrations is a basic input for the application of data management methods. However, not enough official data is available to the public. Although scientific papers report As and F contents at specific sites, a compilation has recently started by a network with participants from the academy, civil organizations and environmental entrepreneurs named “Inventario Nacional de Calidad del Agua (INCA)”. A first task of INCA was to get a comprehensive database of As and F concentrations from papers, thesis, research projects and governmental information to identify those assuring reliable values that were plotted in an interactive map. These actions constitute a starting point to go on with the application of machine learning tools that must include possible predictors of As and F presence like the geology, mineralogy, hydrogeology, etc. in specific zones. A transdisciplinary approach including experts on geochemistry, artificial intelligence, data science, hydrogeology and social scientists, must be created to achieve successful results to be applied by decision makers involved on supplying safe water to the Mexican population.