Filtering the ethics of AI