Does Social Bot Help Socialize? Evidence from a Microblogging Platform.

This article investigates the socializing value of large language model (LLM)-based social bots in public human-bot interactions on social media, focusing on CommentRobot, a platform-owned chatbot deployed on Weibo, a leading Chinese microblogging platform. Empirical analyses reveal that posts receiving bot-generated comments experience significantly higher user engagement—measured by likes and comments—due to the bot's identity combined with high-quality, relevant, and socially expressive content. While the bot's current targeting strategy shows some inefficiencies, policy learning methods can optimize engagement by tailoring bot interactions based on user and post characteristics. However, despite increased engagement at the post level, bot comments primarily stimulate more bot-related posts rather than a broader increase in overall user posting activity, indicating a boundary to the bot's influence on user content creation. The findings contribute to understanding social bots' roles in public online settings and offer practical insights for platforms aiming to enhance user interaction through AI-driven tools.