Your organization has amassed extensive customer information over the years, including addresses and credit card details, which you intend to utilize for building machine learning models on Google Cloud. However, you're apprehensive about potential data leakage into the models, a concern shared by management due to potential reputational damage from direct personal data leaks. How can you address these data security concerns effectively?