[Paper Review] Intelligent social bots uncover the link between user preference and diversity of news consumption

Yong Ki Min, Tingjun Jiang+3 more|arXiv (Cornell University)|Jul 5, 2019

Opinion Dynamics and Social Influence51 references4 citations

TL;DR

This study uses intelligent social bots to simulate user preferences on Weibo, revealing that entertainment-content preference drives stronger information polarization (80% same-type content) than sci-tech preference, due to distinct network sub-structures—star-like for entertainment and clustering for sci-tech—highlighting preference as a key amplifier of polarization beyond self-selection or pre-selection mechanisms.

ABSTRACT

The boom of online social media and microblogging platforms has rapidly alter the way we consume news and exchange opinions. Even though considerable efforts try to recommend various contents to users, loss of information diversity and the polarization of interest groups are still an enormous challenge for industry and academia. Here, we take advantage of benign social bots to design a controlled experiment on Weibo (the largest microblogging platform in China). These software bots can exhibit human-like behavior (e.g., preferring particular content) and simulate the formation of personal social networks and news consumption under two well-accepted sociological hypotheses (i.e., homophily and triadic closure). We deployed 68 bots to Weibo, and each bot ran for at least 2 months and followed 100 to 120 accounts. In total, we observed 5,318 users and recorded about 630,000 messages exposed to these bots. Our results show, even with the same selection behaviors, bots preferring entertainment content are more likely to form polarized communities with their peers, in which about 80\% of the information they consume is of the same type, which is a significant difference for bots preferring sci-tech content. The result suggests that users preference played a more crucial role in limiting themselves access to diverse content by compared with the two well-known drivers (self-selection and pre-selection). Furthermore, our results reveal an ingenious connection between specific content and its propagating sub-structures in the same social network. In the Weibo network, entertainment news favors a unidirectional star-like sub-structure, while sci-tech news spreads on a bidirectional clustering sub-structure. This connection can amplify the diversity effect of user preference. The discovery may have important implications for diffusion dynamics study and recommendation system design.

Motivation & Objective

• To investigate how user preferences influence news consumption diversity and polarization on social media platforms.
• To examine the role of content type (entertainment vs. sci-tech) in shaping the structure of personal social networks.
• To explore the coevolution of user preference, network structure, and information diffusion in a controlled environment.
• To assess whether preference-driven behavior amplifies polarization beyond known mechanisms like self-selection and pre-selection.
• To provide empirical insights for improving recommendation systems and understanding information diffusion dynamics.

Proposed method

• Deployed 68 intelligent social bots on Weibo, each simulating human-like behavior by following 100–120 accounts for at least two months.
• Used text classification algorithms to enable bots to select and follow accounts based on content preference (entertainment or sci-tech).
• Simulated social network evolution using two sociological principles: homophily (like-with-like) and triadic closure (mutual connections).
• Tracked and recorded 630,000 messages across 5,318 users to analyze information exposure and network structure.
• Classified content types using manual and automated text classification to ensure preference fidelity in bot behavior.
• Analyzed local network sub-structures (e.g., star-like, bidirectional clustering) to link content type with propagation patterns.

Experimental results

Research questions

• RQ1 How does user preference for entertainment versus sci-tech content affect the diversity of news consumption in social networks?
• RQ2 To what extent does content preference shape the structural formation of personal social networks on Weibo?
• RQ3 How do different network sub-structures (e.g., star-like vs. clustering) facilitate the spread of specific content types?
• RQ4 Does preference-driven behavior amplify polarization more than self-selection or pre-selection mechanisms?
• RQ5 What is the coevolutionary relationship between user preference, network structure, and information diffusion?

Key findings

• Bots with entertainment preferences formed polarized communities where approximately 80% of consumed content was of the same type, indicating strong information homogeneity.
• Bots with sci-tech preferences exhibited significantly higher content diversity, with less than 50% of content being of the same type.
• Entertainment news preferentially spread through unidirectional star-like sub-structures, enabling efficient one-way broadcasting.
• Sci-tech news spread more effectively through bidirectional clustering sub-structures, supporting mutual reinforcement and information exchange.
• User preference emerged as a more critical factor in limiting content diversity than self-selection or pre-selection mechanisms.
• The coevolution of preference and network structure amplifies polarization, with content type directly shaping the propagation sub-structure within the same social network.