ChatGPT and the Social Appropriation of AI on Discord

Abstract

This study investigates the social appropriation processes of generative AI technologies, focusing on the creation, sale, and sharing of prompts within digital platforms. We aim to transcend traditional socio-technical narratives and imaginaries, examining how users adapt and integrate these technologies into their everyday lives from a grassroots perspective.

Our research centers on an unofficial ChatGPT server on the Discord platform, specifically analyzing three channels related to prompt creation: “Prompt Engineering,” “Project,” and “Image.” Employing a digital ethnography approach, we utilized both quantitative and qualitative methods. We collected 288 posts from the Prompt Engineering channel, 205 from the Project channel, and 4,063 from the Image channel (including 634 posts and 3,429 prompts) using a Python script with Beautiful Soup 4 and Selenium libraries. Additionally, we engaged with users through participant observation and direct interaction, responding to posts about prompt sales and purchases.

Our findings reveal a complex landscape where users adapt AI technologies according to their diverse needs and objectives. We identified four distinct user types: AI Creators, Startuppers, Technology Experts, and AI Curious users. AI Creators utilize generative AI to produce engaging content for large social media audiences; Startuppers leverage AI for financial gain by developing applications like automated marketing tools and chatbots; Technology Experts, including programmers and data scientists, enhance AI functionality through plugin development and share their expertise in prompt creation; and AI Curious users explore AI for emotional and psychological support through personalized interactions.

Furthermore, our study highlights the dynamic interplay of technical and interpersonal skills in the development of a prompt market. Users navigate both public and private channels across multiple platforms to buy and sell prompts, revealing a marketplace where vendors facilitate deeper AI integration into users' routines through pre-configured and customized solutions. This process reflects a form of social appropriation intersecting with economic and emotional dimensions.

Overall, our research demonstrates how the interaction among various user categories and their objectives, ranging from reputation-building and technical skill enhancement to economic and emotional needs, forms the foundation of bottom-up social appropriation practices of AI on the Discord platform.