Author(s):
NeedThe development and rapid diffusion of generative artificial intelligence (AI) represents a great disruption to writing education. Writing and communication are crucial to engineers, taking up more than half their working hours. However, too few engineers have the writing skills requisite for today’s information society. Within this context, new generative artificial intelligence (AI) tools pose both opportunities and challenges for helping engineering students become better writers. On the one hand, these tools provide powerful means of scaffolding the development and practice of writing. They can help ask the right questions, point people toward sources, offer suggestions for getting started, provide model texts, and give coherent feedback on student writing. On other hand, these tools can quickly generate coherent texts on almost any topic and many students–and especially technically-savvy, writing-apprehensive engineering students–may be tempted to unethically use these tools to write texts that they themselves have been assigned, shortchanging students’ own development as writers.Guiding QuestionsOur study seeks to tackle this issue by both studying the integration of AI writing tools in an undergraduate engineering writing course and creating open-source products to help our instructors, and eventually others, integrate AI writing tools into their own courses. We seek to understand the best practices for integrating AI writing tools into engineering writing instruction and learn how such integration impacts engineering students’ development as writers. OutcomesWe are currently conducting design-based implementation research on the incorporation of AI writing tools for scientific communication, exploring best practices for teaching and learning the use of AI writing tools in scientific communication. Thus, our current “outcomes” are threads that we are seeing as we pilot the initial tool and resources in two sections of one instructor’s course. Our initial instructional framework, that students would need to understand the tools’ functions, strengths, and weaknesses; access a range of tools; learn prompt the tools; understand the critical importance of corroborating the accuracy and truthfulness of AI output; and practice incorporating AI-generated texts into their own writing ethically and effectively have held up well. In this poster we will present the key lessons learned so far for both instructors and students. Before instructors (or students) turn to the AI, they should think first on their own. While working with the AI, students should ask questions, push back, have the AI expand or clarify as desired, we refer to this as “being the boss” of the AI. Finally, after the writing process is complete, writers need to reflect on the process and outcome, and particularly the role of the AI. Writers need to determine what worked for them and what did not, as well as what they might do differently next time.Broader ImpactsThe lessons learned so far have been shared with two sections of engineering writing students (n=37) and 4 writing instructors, as well as numerous others as the team actively disseminates our work. As we continue this work, we will be scaling up to 8 sections of students in spring quarter (n=160).
Coauthors
Waverly Tseng, University of California, Irvine, CA; Beth Harnick-Shapiro, University of California, Irvine, CA; Daniel Ritchie, University of California, Irvine, CA; Mark Warschauer, University of California, Irvine, CA
