Making ChatGPT Work For Me

Abstract

Increasingly, work happens through human collaboration with generative AI (e.g., ChatGPT). In this paper, we present a qualitative study of this collaboration for real-life work tasks. We focus our study on US K12 public school teachers (𝑁 = 24) who regularly design and complete text-generation tasks such as creating quizzes, slide decks, word problems, reading passages, lesson plans, classroom activities, and projects. In one-on-one video- and audio-recorded virtual sessions, we observe each teacher using ChatGPT-4 for work tasks of their choosing for 15 minutes, then debrief their experience. Analyzing 201 prompts inputted by the 24 teachers, we uncover four main modes with which the teachers request support from ChatGPT: (1) make for me (55% of prompts), (2) find for me (15%), (3) jump-start for me (10.5%), and (4) iterate with me (15.5%). The first three modes (make, find, and jump-start) are often requests of generative AI to do something, whereas the fourth mode (iterate) is a request of generative AI to think. In a follow-up survey of the same 24 teachers, most report using multiple modes for their work, but infrequently. Our study contributes new data and knowledge about how teachers are coming to understand whether and how to integrate generative AI into their teaching preparation routines.

Keywords: K12 education, generative AI, human-computer interaction