ABSTRACT VIEW
Abstract NUM 1817

ADDRESSING THE CHATBOT: RESULTS OF A SYSTEMATIC REVIEW OF INSTRUCTOR CONCERNS ABOUT GENERATIVE AI AND CURRICULUM DEVELOPMENT
M. Trice, D. Larson, O. Szabo, K. Parsons, S. Bates
Massachusetts Institute of Technology (UNITED STATES)
We propose an oral panel session of five instructors discussing the incorporation of generative AI assignments into their writing and communication curriculum. The panel reflects upon how these activities were developed and then shared with a department of 40 communication instructors at a private research-focused university in the United States. The panel reviews the development, procedures, and department feedback from these assignments. The assignments were developed to help students engage with ethics, process, and practical applications of generative AI, while also exploring the evolving limits and affordances of the technology.

We first inform our panel by exploring a two-year survey about instructors' concerns regarding generative AI gathered from Spring 2024 through Spring 2025. We highlight instructors' concerns about a lack of resources to fully understand generative AI and an inability to fully revamp their curriculum to modernize assignments for generative AI. From these surveys, we then explain our process for developing training sessions for instructors and developing a range of assignments to meet a wide array of needs, including AI-skeptical and AI-embracing instructors. We then analyze survey feedback from the instructional sections to evaluate how assignments matched instructor needs as expressed in the multi-term survey.

Deliverables from this study include the initial multi-year survey of instructor concerns about GenAI, a description of all five assignments generated, and session feedback surveys from the initial assignment demonstrations. Brief descriptions of the assignments are provided below.

In our first activity, we compare bios of instructors generated by the free and paid versions of ChatGPT to compare outputs and demonstrate the nature of accuracy issues with generative AI.

For the second activity, participants work with generative AI to craft summies of an assigned text. Rhetorical, rather than solely descriptive, choices are emphasized so that the capabilities of genAI can be assessed beyond mimicking general rhetorical moves in genre conventions.

In another activity, students use generative AI to improve grammar in definition writing. They evaluate AI and peer feedback for tone, style, and correctness. The activity includes training, drafting, peer review, and revision, enhancing grammar awareness and AI-assisted writing skills.

Another activity aims to raise students’ awareness of the ethical issue of intellectual property theft by generative AI. The activity begins with a discussion of music piracy, which is then compared to generative AI’s theft of copyrighted images. The instructor shows students a currently pending class action lawsuit related to this issue. The activity also includes a conversation about citation practices when using generative AI to “create” images.

Keywords: Generative AI, assignment design, education.

Event: ICERI2025
Session: AI in Academic Writing
Session time: Tuesday, 11th of November from 08:45 to 10:00
Session type: ORAL