Paper Type
Short
Paper Number
PACIS2025-1317
Description
As retail investor users scale on social media platforms, maintaining the financial informativeness of crowd-generated content becomes difficult. While automated content moderation is widely used in platform governance, its effectiveness for specialized knowledge domains remains underexplored. This study examines how dramatic community expansion influences bot moderation effectiveness in online investment communities by analyzing the GameStop event on Reddit's r/WallStreetBets. Using 1.7 million posts and 419,000 bot moderation comments from 2020-2021, we find that different bot moderations yield varying outcomes, depending on their alignment with the wisdom of crowds boundary conditions. Image-related moderation becomes more effective in the community shift, while humor-embedded moderation grows less effective. Media-related moderation shows different time-dependent effects: positive for short-term returns but negative for longer-term predictions. These findings contribute to platform governance literature on how bot moderation maintains financial information quality during rapid community expansion and provide practical implications for social media platform operations.
Recommended Citation
Hu, Shenghui and Leung, Alvin, "Bot Moderation Dynamics in Online Investment Communities" (2025). PACIS 2025 Proceedings. 16.
https://aisel.aisnet.org/pacis2025/sm_digcollab/sm_digcollab/16
Bot Moderation Dynamics in Online Investment Communities
As retail investor users scale on social media platforms, maintaining the financial informativeness of crowd-generated content becomes difficult. While automated content moderation is widely used in platform governance, its effectiveness for specialized knowledge domains remains underexplored. This study examines how dramatic community expansion influences bot moderation effectiveness in online investment communities by analyzing the GameStop event on Reddit's r/WallStreetBets. Using 1.7 million posts and 419,000 bot moderation comments from 2020-2021, we find that different bot moderations yield varying outcomes, depending on their alignment with the wisdom of crowds boundary conditions. Image-related moderation becomes more effective in the community shift, while humor-embedded moderation grows less effective. Media-related moderation shows different time-dependent effects: positive for short-term returns but negative for longer-term predictions. These findings contribute to platform governance literature on how bot moderation maintains financial information quality during rapid community expansion and provide practical implications for social media platform operations.
Comments
Social