The main cause of bias in AI is the training data. Often, some groups are overrepresented, and others are under-represented. When we have little attention on how data is collected, processed, and organized there will be a lack of geodiversity, which incidentally produces data that leads to gender, ethnic and cultural biases.