"Ronaldo's a poser!": How the Use of Generative AI Shapes Debates in Online Forums

TL;DR

The paper addresses how Generative AI shapes online argumentation in polarized sports forum debates by studying ChatGPT assisted Messi vs Ronaldo discussions. Using a two part design with turn based and free debates, plus semi structured interviews, the authors analyze posts, AI transcripts and participant reflections to map AI mediated argument practices. Key findings show participants seek aggressive AI prompts, AI co writing yields similar posts, and new participants shift usage toward human human engagement while AI assisted synthesis persists. The work highlights practical implications for visualizing logical constructs and building intent based AI collaborators, while noting limitations in generalizability and ecological validity and suggesting directions for future research in AI mediated online discourse.

Abstract

Online debates can enhance critical thinking but may escalate into hostile attacks. As humans are increasingly reliant on Generative AI (GenAI) in writing tasks, we need to understand how people utilize GenAI in online debates. To examine the patterns of writing behavior while making arguments with GenAI, we created an online forum for soccer fans to engage in turn-based and free debates in a post format with the assistance of ChatGPT, arguing on the topic of "Messi vs Ronaldo". After 13 sessions of two-part study and semi-structured interviews with 39 participants, we conducted content and thematic analyses to integrate insights from interview transcripts, ChatGPT records, and forum posts. We found that participants prompted ChatGPT for aggressive responses, created posts with similar content and logical fallacies, and sacrificed the use of ChatGPT for better human-human communication. This work uncovers how polarized forum members work with GenAI to engage in debates online.

Figures (9)

• Figure 1: The study consists of two parts: Part 1: One-on-One Turn-Based Debate and Part 2: Three-Person Free Debate. In Part 1, two participants, Alice and Bob, with opposing stances, engaged in a turn-based debate. In Part 2, a third participant, Carol, who shares the same stance as one of the original participants, joined the ongoing debate. Image credit: ESPN FC.
• Figure 2: One participant (P11) attempted to prompt ChatGPT to generate a poem containing insulting jokes. ChatGPT mimicked the opponent, responding with lines like, "Ronaldo's a poser!" and "Messi's the greatest ...".
• Figure 3: Participants tried to introduce aggressive content by themselves when they were not satisfied with the response of ChatGPT after several times re-prompting. This figure shows an example of P3.
• Figure 4: Participants repeated similar content in their posts several times, reflecting their reliance on the response of ChatGPT. This figure shows an example of P16 and P17.
• Figure 5: Participants shared similar content in their posts, regardless of whether they were opponents (Example 1) or teammates (Example 2). By quoting other members' posts to prompt ChatGPT and using the information generated, their posts contained similar content.