Xiao Tan
Utah State University
This activity is designed to help students understand the concept of “rhetorical reading” (Haas & Flower, 1988) by juxtaposing their human reading and writing experience with the AI-assisted experience. Students become more aware of the rhetorical situation when it is disrupted by the intervention of AI technologies. At the end of the activity, students will gain first-hand experience of reading rhetorically, develop understanding of the limitations and affordances of AI tools, and critically examine the loss and gains in using AI technologies in academic settings.
Learning Goals
Original Assignment Context: A theme-based first-year writing course focusing on AI and writing
Materials Needed
Time Frame: One class session (~75 minutes)
Overview
This activity is designed to help students understand and experience the concept of “rhetorical reading” (Haas & Flower, 1988). It has been taught one time in one of my first-year composition courses in Spring 2024 and has yielded good results. This activity takes only one 75-minute class session and involves students in several writing, discussion, and collaboration tasks. In this activity, students juxtapose a summary they wrote with a summary generated by ChatPDF or similar AI-powered tools and discuss how the simple process of “reading” and “summarizing” could be mediated, shaped, or even disrupted by the involvement of AI tools. This activity also engages students in the critical discussion of whether human reading experiences could be replaced by the use of AI, and if so, what is gained and lost in the process. Students showed high engagement during the in-class portion of this activity, critically evaluating the (in)accuracy of AI-summarized texts and identifying their personal connections with the readings. In a follow-up discussion activity on Canvas, students also extended the conversation to the philosophical debate of whether writing style can be detached from the content. Overall, I find this activity an engaging and intriguing one that could be adapted to courses at various levels. Due to the low knowledge threshold of using AI tools, students at different familiarity levels could benefit from the experience of reading intentionally with AI and exchanging ideas with peers.
Before the class, the class is divided into Group A and Group B. Students in Group A are assigned to read “Meta-Writing: AI and Writing” by Aimée Morrison (2023), and Group B is assigned to read “Defining Moments, Definitive Programs, and the Continued Erasure of Missing People” by Alfred Owusu-Ansah (2023). The two articles were published in the same special issue on AI in Composition Studies and were chosen to provide students with different but intriguing perspectives of AI and writing. Moreover, these two articles have distinct writing styles that evoke different reading experiences. Owusu-Ansah’s piece aligns closer with the prototype of a research article in structure and language style, while in Morrison’s article the author adopts a narrative style akin to the stream of consciousness and shares her free-flowing thoughts. Writing instructors may also choose other reading materials that align with the topic or theme of their courses. In addition to reading the assigned articles, all students read “Rhetorical Reading Strategies and the Construction of Meaning” by Christina Haas and Linda Flower (1988).
At the beginning of the class, I introduce ChatPDF to the whole class and demonstrate the basic functions of ChatPDF. This platform can process uploaded files and help summarize a text, locate information, and answer questions based on the text. After the demonstration, students are given 30 minutes to summarize the reading assigned to them and prompt ChatPDF to generate a summary of the similar length for the other article. The prompt for this writing task is open-ended, asking students to summarize the main arguments, author’s intention, and supporting ideas.
Following the individual writing task, each student from Group A is paired up with a student from Group B. Working in pairs, students compare their summaries (one written by humans and one generated by the chatbot) for the same article and discuss textual and rhetorical similarities and differences. They also compare their reading and writing experience, as well as their experience of using ChatPDF to summarize a text.
In the last 15 minutes of the class session, the whole class shares their experiences and observations from the writing tasks and pair work. As homework, students reflect on the in-class activity and addresse the following questions on Canvas:
1. What are the major differences between the AI-assisted summary and your summary? How do the reading processes differ?
2. When reading the article, how did you use the rhetorical strategies described in Haas and Flower (1988)?
3. What is gained and lost when reading with the assistance of AI? When is it appropriate and ethical to use ChatPDF to help with reading?
Overall, the activity was well received, as evident in the active engagement and critical thinking demonstrated by students. Juxtaposing the two summaries brings to students’ attention the “rhetorical aspect” of reading. Many students mentioned in class discussion and written reflections that when reading and summarizing they tended to draw on their past experiences and even emotional responses to process the textual meaning, all of which seemed to be lost in AI summaries. Some students pointed out that in their own summaries they included or focused on information that resonated most strongly with their personal experiences, while the chatbot devoted equal attention to each section. On the same note, several students also discussed the limitations of ChatPDF, highlighting the lack of semantic accuracy and possible distortion of meaning in AI-generated summaries. Amidst the critical comments on the chatbot’s limitations, some students noticed that the writing style and structure of the original article could have an impact on AI-generated outcomes. For example, Owusu-Ansah’s piece seemed to be processed more accurately than Morrison’s piece because the former follows a structure typical to research paper. Moreover, one student challenged the idea that AI summaries are always ineffective by demonstrating her strategies of getting more accurate responses. Instead of entering the simple command of “summarizing the article,” the student asked ChatPDF to summarize each section and then synthesized the summaries herself.
While the class discussion centered mostly around the concept of “rhetorical reading,” students also extended the conversion into the debate about content and form. One student criticized Morrison’s way of making arguments, commenting that the author only gets to the main point at the end of writing. Another student refuted this comment, arguing that how arguments are presented is essentially a part of the argument. In other words, the “stream of consciousness” writing style in Morrison’s article, which is an intentional rhetorical choice, illustrates her argument that “the advantage that human writers have–beyond our crystallized knowledge, contextual understanding, desire to connect, and gift for linguistic and syntactic innovation–is that we seek, produce, and discern meaning” (p. 160). This activity brings forth the hidden aspect of argument-making and encourages students to analyze both what is presented and how it is presented.
I would like to acknowledge the Thompson Writing Program at Duke University where this activity was first implemented.