With the proliferation of easy-to-use generative AI tools such as ChatGPT, students in college composition courses can and do take advantage of such technology to assist with essay writing. While a growing body of research does not see using these tools as a problem in itself, institutional, departmental, and instructor policies about generative AI usage to assist with essay writing are often confusing due to their language. Furthermore, the ethics of using ChatGPT to assist with the writing process is often highlighted as important by educational policymakers, but the spectrum of ethicality for different applications of generative AI often goes unaddressed. In this paper, I propose a taxonomy of five categories of AI usage in the composition classroom, each with a corresponding level of ethicality. Largescale adoption of this taxonomy would not only standardize language for policymakers, instructors, and researchers, but it would help inform both teachers and students about different applications of generative AI tools in the composition classroom beyond simple copy and paste of AI generated paragraphs.