On conflicts between ethical and logical principles in artificial intelligence