Abstract
Artificial intelligence (AI) in general and specifically large language models (LLMs), especially ChatGPT, have generated great interest in education. Faculty now need to consider how we can better teach students to use these tools, both ethically and functionally. While educational institutions in the last 20 years or so have made strides in improving students’ digital literacy skills, educators now are on a precipice of change that requires us to teach students a fundamentally different kind of skillset, one grounded in AI-enhanced literacy. We need to teach students how to interact with these tools effectively and efficiently. LLMs in particular require a skillset known as “prompt engineering,” the ability to interact with these models through written questions and commands. The old adage “garbage in, garbage out” applies here because without understanding how to interact effectively with LLMs, students will generate less-useful output. My paper focuses on a class I developed in conversation design for primarily undergraduate English and Communications majors. The purpose of this class is to present a scaffolded approach to prompt engineering by first teaching students conversation design skills in the context of creating their own rule-based chatbot to designing skills for an AI-enhanced voice assistant such as Alexa and Google Assistant. Finally, with these skills in hand, students are introduced to prompt engineering with ChatGPT. The purpose of this approach is to provide students, most of whom had no technical background, with an understanding of how to communicate with AI and develop skills for the new workplace.
Presenters
Mark MabritoProfessor, English Department, Purdue University Northwest, Indiana, United States
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
Artificial Intelligence, Conversation Design, Prompt Engineering, ChatGPT, Writing