Your Name and Title:
Jon Ippolito, Professor of New Media and Director of Digital Curation
School, Library, or Organization Name:
School of Computing and Information Science, University of Maine
Co-Presenter Name(s):
Gregory Nelson, Troy Schotter
Area of the World from Which You Will Present:
North America
Language in Which You Will Present:
English
Target Audience(s):
Educators and researchers looking for pedagogical models and data on AI in education across media, tools, and genres.
Short Session Description (one line):
Learn the results of a semester-long study in which 50 students rated their experience with AI versus more traditional tools, from writing essays to coding game avatars to creating 2d, 3d, and time-based media.
Full Session Description (as long as you would like):
Numerous studies have focused on the impact of ChatGPT and its ilk on student essays, yet writing is far from the only educational task disrupted by generative AI tools. This presentation widens the lens to ask how AI tools of all kinds delight and confound the incoming cohort of college students, from the ability to create art, code, and text to confidence about personal capabilities and future career prospects.
Appraising the full range of AI output is increasingly important for the post-ChatGPT classroom. Given the fallibility of AI writing detectors, multi-modal projects that combine text with other media have become versatile replacements for the term paper. Asking students to generate images with tools like Stable Diffusion and Midjourney, meanwhile, can reveal inner workings and challenges of large language models that are hidden by ChatGPT's oracular interface.
Last fall University of Maine Computer Science professor Greg Nelson, graduate student Troy Schotter, and New Media professor Jon Ippolito designed an entire Introduction to New Media course to compare student experiences performing tasks with pre- and post-generative AI approaches. In the first half of the semester, 50 students wrote texts, programmed software, and created media using familiar digital tools; in the second half, they performed the same tasks using generative AI.
In Week 7, for example, students revisited their first writing assignment from Week 1--identifying winners and losers in a field disrupted by a contemporary technology--but were required to lean on GPT-4 for help brainstorming, drafting, and polishing a similar essay with a different choice of field and technology. Students then wrote a detailed comparison of the two drafts. As another example, in Week 10 they did a similar comparison of Photoshop and Stable Diffusion for illustrating a hoax, from the Loch Ness Monster to JFK as a Masonic figurehead.
The class repeated this process for designing logos, coding game avatars, creating illustrations, writing stories, recording soundscapes, and even grading homework. This conversation will review preliminary findings of this ambitious experiment, including how success varied by assignment and medium, the difference between evaluation by chatbots versus human TAs, and how the experience impacted student confidence overall.
Websites / URLs Associated with Your Session:
AI versus old-school creativity: a 50-student, semester-long showdown
An illustrated summary of preliminary findings
NMD 100 Introduction to New Media
The course in which the study took place
The University of Maine's AI learning toolkit
Replies