Litcius/Paper detail

Generative AI in first-year writing: An early analysis of affordances, limitations, and a framework for the future

Robert Cummings, Stephen M. Monroe, Marc Watkins

2024Computers & composition/Computers and composition65 citationsDOIOpen Access PDF

Abstract

Our First-year Writing program began intentional student engagements with generative AI in the fall of 2022. We developed assignments for brainstorming research questions, writing counterarguments, and editing assistance using the AI tools Elicit, Fermat, and Wordtune. Students felt that the tools were helpful for finding ideas to get started with writing, to find sources once they had started writing, and to get help with counterarguments and alternate word choices. But when given the choice to use the assistants or not, most declined. Generative AI at this stage is unreliable, and many students found the tradeoff in reviewing AI suggestions to be too time consuming. And many students expressed a preference for continuing to develop their own voices through writing. Our experience in engaging AI led to the creation of the DEER praxis, which emphasizes defined engagements with AI tools for specific purposes, and generous use of reflection.

Topics & Concepts

AffordanceGenerative grammarComputer sciencePsychologyCognitive scienceLinguisticsMathematics educationHuman–computer interactionArtificial intelligencePhilosophyIntelligent Tutoring Systems and Adaptive LearningTopic ModelingInnovative Teaching and Learning Methods
Generative AI in first-year writing: An early analysis of affordances, limitations, and a framework for the future | Litcius