Building Generative AI Literacy During the Writing Process

Submitter: Kyle Jensen, Arizona State U

——————————————————

The experiment:

In Spring 2023, I taught a research methods course in our Rhetorics, Writing, and Literacies undergraduate program. This course introduces students to advanced research methods such as rhetorical analysis, historiographical analysis, archival analysis, narrative analysis, and discourse analysis. I organized course readings around the histories, theories, and technologies associated with generative artificial intelligence. I wanted students to conceptualize generative AI research as a problem solving process where researchers implement different methods and measure the results. To measure student learning, I asked students to identify a topic within generative AI research that they found compelling, define an associated problem using methods we discussed in class, and negotiate the problem by tracking the competing perspectives that engage it. Students used generative AI applications while they invented, drafted, and revised their written assignments. Doing so allowed them to apply the research methods they were studying. It also allowed them to develop critical literacy with generative AI applications. We focused on ChatGPT developed by OpenAI and Wordtune developed by AI21 Labs.

Results:

Students expressed two concerns about generative AI applications at the beginning of the course. First, they were unsure about using generative AI to complete their assignments as doing so may constitute an academic integrity violation. Second, they were unsure if generative AI applications would co-opt their voice during the writing process. We addressed both concerns by testing different types of generative AI applications. By using applications that were not prompt-based, whole text generators, students learned that generative AI and ChatGPT are not coextensive with one another. As a result, they discovered that different applications were valuable at different stages of the writing process. For example, they found that prompt-based applications such as ChatGPT were especially helpful as an invention tool. But as they moved into drafting, revision, and reflection, they valued generative AI applications such as Wordtune that generated multiple possibilities from what they had written. In their final course reflections, they explained that such applications encouraged them to make thoughtful decisions about their final revisions.

Relevant resources:

• https://www.washingtonpost.com/technology/2023/06/28/ai-gen-z-work/
• https://news.asu.edu/20230523-asu-students-explore-ways-embrace-ai-tools-writing-course

Contact: dr.kjensen[AT]asu[DOT]edu