Computers and Composition最新文献

The importance of peer review practice in writing courses has been strongly supported by pedagogical research. Adopting a mixed-methods approach, this study investigated three peer review modes in an undergraduate academic writing course through the lens of students’ writing performances and perceptions. The three modes are (i) face-to-face peer review (F2F), (ii) anonymous computer-mediated peer review (CMPR), and (iii) blended peer review (a blend of F2F and anonymous CMPR). Three classes enrolled in an academic writing course participated in this study. Students’ assignments were collected to analyze their writing performances. Focus group discussions (FGDs) were administered to investigate students’ perceptions of the peer review modes, including their perceived usefulness of the feedback and the review processes. The findings show that the students’ writing performances significantly improved after the peer review session in all three peer review modes, with the anonymous CMPR and the blended mode showing stronger effectiveness as compared to the F2F mode. The participants generally preferred the blended mode, which addresses the limitations of both F2F and anonymous CMPR by leveraging the merits of both. We propose the use of the blended peer review mode to accommodate different learning needs and maximize the effectiveness of peer review practice.

Research approaches that emphasize embodied practice and value the idiosyncratic uptake of writing technologies should be central to how writing researchers process the early moment of generative AI's availability to public audiences. Based on a qualitative analysis of 35 publicly available videos depicting the use of ChatGPT and Bing, the study offers a framework of descriptive codes that identify practices early adopters enact when integrating these technologies into research writing processes. The research contributes three key categories of practice that describe research writers’ interaction with generative AI across research design, writing research genres, and proofreading and editing: requesting, evaluating, and refining. This study is significant for providing an early descriptive analysis of the uptake of ChatGPT in research writing, while also identifying how disparate uses of generative AI technologies emerge from conflicting beliefs about writing, research, and invention. In particular, the study describes how experts in writing and research portray uses of and attitudes toward these technologies that often differ from students who are learning to research and write in their respective fields.

Ethical frameworks for text generators (TGs) in education are generally concerned with personalized instruction, a dependency on data, biases in training data, academic integrity, and lack of creativity from students. While broad-level, institutional guidelines provide value in understanding the ethical dimensions of artificial intelligence (AI) for the classroom, there is a need for a more ecological understanding of how AI ethics might be constructed locally, one that takes into account the negotiation of AI between teacher and student. This article investigates how an educational ethical framework for AI use emerges through a qualitative case study of one composition student's interaction with and understanding of using ChatGPT as a type of writing partner. Analysis of interview data and student logs uncover what we term an emergent “local ethic” – a framework that is capable of exploring unique ethical considerations, values, and norms that develop at the most foundational unit of higher education – the individual classroom. Our framework is meant to provide a heuristic for other writing teacher-scholars as they interrogate issues related to pedagogy, student criticality, agency, reliability, and access within the context of powerful AI systems.

Compositionists can train students to rhetorically approach AI as a writing assistant—a metaphor that can help students maintain authority in dialogue with AI. Such an approach allows students to control the input and output from generative AI. Compositionists can teach students how to input ethical information and recognize harmful output. The students are then encouraged to converse with the AI in dialogue, sending messages back and forth until the AI generates writing that is acceptable in ethics and content. This article provides a social justice heuristic for compositions to follow in their training of students.

This article offers Rhetorical Load Sharing as a theoretical framework for placing texts on a collaborative authorship spectrum spanning from human-authored text to synthetic text. It poses human-in-the-loop writing as a baseline ethical AI collaborative writing workflow that avoids offloading the entire rhetorical load to generative AI tools and argues that machine-in-the-loop writing, in which human collaborators retain majority of the rhetorical load, is an ideal AI collaborative writing model that is suitable for the technical and professional communication classroom.

We share our experiences working with large-language model generative AI for a full semester in a professional writing course, integrating it into all projects. We discuss how we adapted our teaching, learning, and writing to using (or purposefully not using) AI. Issues we discuss include balancing integration of AI to avoid potential overreliance, the importance of centering authorial agency and decision-making, negotiating grading and evaluation, the benefits and drawbacks of AI throughout the writing process, and the relationships we build or could build with AI. We close with recommendations for faculty and students.

This study introduces online advertising platforms as digital composing tools where persuasive rhetoric encourages users to follow links and take action on landing pages. It frames these platforms as digital spaces where human actors work alongside non-human AI agents (Duin & Pedersen, 2021 & 2023) and where rhetorical agency emerges through the activity of machine learning and artificial intelligence. It theorizes a (human) user-centered approach to composing digital ads in digital advertising platforms built around Walton, Moore & Jones’ (2019) framework of positionality, position, and power. It provides guidance for technical and professional writers in placing human users at the center of an abstracted, algorithm-driven partnership where generative AI appears poised to wrest power from both composers and users.