Are all machine learning models bad? Surely there could be some machine learning models that are so simplistic in nature that they could be considered "harmless

Are all machine learning models bad? Surely there could be some machine learning models that are so simplistic in nature that they could be considered "harmless." It seems like the purpose and scope of the model is not considered by most people when they are debating over AI.