Curricular Implementation of Large Language Models to Augment the Design Process and Technical Communication in an Introductory Mechanical Engineering Course

Abstract

Abstract The rapidly increasing quality and ease of access to highly sophisticated Large Language Models (LLMs) (e.g., ChatGPT, Claude, CoPilot, etc...) has created a situation of near ubiquity in college campuses across all disciplines, eliciting responses from administrators and educators ranging from exuberant to apocalyptic. Regardless of your position on the topic, it should be acknowledged that bans on the use of such technology are practically unenforceable and oppose the development of effective pedagogical practices centered on their use. Instead, educators should introduce, reinforce, and encourage responsible and critical use of Generative AI tools. Considered through the lens of Bloom’s taxonomy, students should be encouraged to create prompts that fall between ‘Recall’ (questions more appropriate for a search engine) and ‘Create’ (replacing their own critical thinking). This work will focus on the implementation of LLM prompting exercises in an introductory engineering course. The course in question introduces students to the engineering design process, technical communication, and computer-aided design. The course policy has evolved its approach to Generative AI from one of tacit encouragement (i.e., Generative AI is allowed unless specifically forbidden) to one of active encouragement (assignments dictate the use, documentation, and assessment of Generative AI). This work will describe efforts to direct the use of Generative AI in individual design and technical communication graded assignments through prescribed and/or suggested prompts. At the end of the course, a survey was conducted with all students to assess their previous and current comfort with LLM interaction, as well as the perceived efficacy of the technology and students’ perceptions regarding the responsible use of generative AI, both as students and future professionals.

Ähnliche Arbeiten