TextGenEd: Teaching with Text Generation Technologies
CALL FOR PROPOSALS: TextGenEd: Teaching with Text Generation Technologies
Annette Vee, Assoc. Prof. of English and Dir. of Composition, University of Pittsburgh
Tim Laquintano, Assoc. Prof. of English and Dir. of College Writing Program, Lafayette College
Carly Schnitzler, Ph.D. Candidate, University of North Carolina-Chapel Hill
We invite college instructors working with text generation technologies (e.g., GPT-3, Markov models, Tracery) to submit relevant classroom assignments and activities to an open access edited collection. This collection will serve as a teaching resource, bringing together approaches to AI and procedural text generation across the curriculum, including: composition, education, literature, digital humanities, linguistics, computer science, creative writing, technical communication, computational poetics, and writing in the disciplines. Assignments (at the undergraduate or graduate level) might encourage students to grapple with the technical, ethical, historical, or stylistic aspects of text generation technologies.
Text generation technologies and online services such as GPT-3, Tracery, RiTA, Sudowrite and Jasper.ai have, in the last couple of years, become viable producers of coherent writing. Creative writers are finding use for them in their own writing practices, bloggers are using them for generating content quickly, and students are using the tools to augment their writing processes. Teachers of writing are wrestling with these new tools, their technical and stylistic details, and the ethics of using them. Some teachers are embracing the tools by having students experiment with generating text, fine-tuning language models to particular genres, engineering prompts to produce text, revising generated text, or analyzing generated text output. Our collection asks teachers to submit those assignments and activities, along with their context, results, and the teacher’s insights, and organize them for reuse by other teachers of writing. The goal of the collection is neither to promote generated writing nor to dismiss it, though tech-forward and critical submissions are both welcome. We aim to gather innovative assignments that ask students to contend with the conditions of contemporary writing and help teachers to navigate those conditions as well, even if they might not yet be aware of them.
Contributions could include assignments that ask students to:
Compose with online interfaces such as Sudowrite to brainstorm, produce, or edit text
Analyze the output of LLMs (large language models) such as GPT-3 and compare it to human writing
Probe biases in the output of LLMs and other text generation technologies
Explore the ethical implications of text generation technologies through hands-on work
Draft policies that delineate the ethical uses of text generation technologies in writing education
Trace implementations of text generation technologies in writing software
Use text generation technologies for game design and critique
Do anything else related to text generation technologies
We welcome contributions from across the undergraduate and graduate curriculum—anywhere writing is taught, including in creative, academic, technical, legal, composition, linguistic, historical, and computational contexts. The collection as a whole will attend to a diversity of student and teacher backgrounds and aim to be as accessible as possible.
We are calling for 250-400 word proposals that summarize the assignment, including its format and length requirements, context and implementation, learning goals or outcomes, and the preparation, materials, and skills necessary to complete it for both teachers and students. Authors should each include a 50-75 word bio with their submission. Joint and multiple submissions are welcome. If needed, any citation style (e.g., MLA, APA, Chicago) is welcome for the proposal. Given the fast pace of integration of text generation technologies in writing, we have an accelerated timeline and open access plan for this collection. We plan to share the Table of Contents in early Summer 2023 and have the collection published by August 2023, thus available for adaptation for the academic year 2023-24. We are currently in conversation with an online publishing venue that is committed to this publication timeline and will support open access, peer review, and Creative Commons reuse. To submit, please email your proposal to firstname.lastname@example.org as a .docx attachment by 11:59 ET on December 20, 2022.
Please direct questions to email@example.com .
Tentative timeline to publication
Dec 20, 2022: 250-400 word proposals due to Editors
Jan 15, 2023: Editors response to proposals
Feb 15, 2023: Full submissions due
Mar 15, 2023: Internal reviews due: each submitter (or team) is asked to review two full submissions
Apr 15, 2023: Revisions due
May 15, 2023: Collection out for full review
Jul 15, 2023: Revisions to collection completed
Aug 15, 2023: Publication online
Each submission will eventually include (in this order):
Introduction to the assignment and the teacher's experiment, including description of how many times it's been taught, in what context, with what outcomes (it must have been taught at least once, and we embrace failed assignments!)
Goals and outcomes of the assignment
Materials needed including software/hardware/skills required
Acknowledgments for any assignment, collaborator, curriculum designer, researcher or publication that influenced the assignment
Assignment in a format that could be adapted by others (using a CC-NA license)
Length is flexible for the full submissions, but we encourage brevity for better circulation, review, and adaptation. We suggest introductions be about 1000 words and assignments be about two pages.
All assignments for this collection will be licensed under a Creative Commons Noncommercial Attribution license so that they may be adapted by other teachers.