arXiv:2605.24000v1 Announce Type: new Abstract: Toxicity in online gaming communities remains a persistent challenge, manifesting across genres, platforms, and player interactions. While much research is focused on in-game toxicity, less is known about how toxic behavior varies between gaming communities on streaming platforms. To address this shortcoming, we analyze approximately 20 million chat messages from 4,452 streams, spanning seven game genres on Twitch. We categorize messages according to Twitch's toxicity taxonomy with a pre-trained Large Language Model using zero-shot classification
The proliferation of AI, particularly large language models, allows for unprecedented large-scale analysis of unstructured data like chat logs, making such studies feasible and efficient today.
This study demonstrates how LLMs can be effectively applied to analyze complex social dynamics in vast online communities, offering new tools for content moderation and understanding digital behavior.
The ability to automatically categorize and analyze toxicity across diverse gaming communities with LLMs provides a scalable method for platforms to address persistent issues, moving beyond reactive human moderation.
- · Twitch
- · Gaming platforms
- · AI developers
- · Online communities
- · Toxic users
- · Manual content moderation
Increased understanding of toxicity patterns and variations across online communities.
Development of more sophisticated, AI-driven content moderation tools and policies on streaming platforms.
Potential for AI to shape community norms and communication styles by proactively identifying and mitigating negative interactions.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL