[Paper Review] Threat or Opportunity? - Examining Social Bots in Social Media Crisis Communication

Florian Brachten, Milad Mirbabaie+4 more|arXiv (Cornell University)|Oct 22, 2018

Misinformation and Its Impacts26 references4 citations

TL;DR

This study investigates social bot activity during the 2017 Manchester Bombing using a case study approach, analyzing Twitter data to assess bot behavior in crisis communication. Findings reveal that while bot accounts are numerically scarce, they exhibit high engagement levels, primarily serving benign informational roles rather than spreading disinformation, suggesting bots can be an opportunity rather than a threat in crisis contexts.

ABSTRACT

Crisis situations are characterised by their sudden occurrence and an unclear information situation. In that context, social media platforms have become a highly utilised resource for collective information gathering to fill these gaps. However, there are indications that not only humans, but also social bots are active on these platforms during crisis situations. Although identifying the impact of social bots during extreme events seems to be a highly relevant topic, research remains sparse. To fill this research gap, we started a bigger project in analysing the influence of social bots during crisis situations. As a part of this project, we initially conducted a case study on the Manchester Bombing 2017 and analysed the social bot activity. Our results indicate that mainly benign bots are active during crisis situations. While the quantity of the bot accounts is rather low, their tweet activity indicates a high influence.

Motivation & Objective

• To understand the role of social bots in social media crisis communication during extreme events.
• To investigate whether bots contribute to information dissemination or misinformation during crises.
• To analyze the behavioral patterns and influence of bots during the Manchester Bombing in 2017.
• To assess whether bot activity during crises is predominantly benign or malicious.
• To contribute empirical evidence on bot dynamics in real-world crisis scenarios, addressing a gap in HCI and social media research.

Proposed method

• Conducted a case study focusing on the Manchester Bombing event in May 2017.
• Collected and analyzed Twitter data during the crisis period using automated data collection tools.
• Identified bot accounts using established bot detection heuristics based on behavioral patterns such as tweet frequency, timing, and network structure.
• Classified bot activities into categories such as information sharing, retweeting, and engagement to assess their functional roles.
• Quantified bot activity by measuring tweet volume, engagement rates, and temporal distribution relative to human users.
• Compared bot behavior with human user activity to evaluate influence and information flow dynamics.

Experimental results

Research questions

• RQ1 What types of social bots are active during crisis events like the Manchester Bombing?
• RQ2 To what extent do social bots contribute to information dissemination during crises?
• RQ3 Are social bots primarily involved in spreading misinformation or providing helpful information in crisis contexts?
• RQ4 How does the volume and activity level of bots compare to human users during crisis events?
• RQ5 What is the perceived impact of bot activity on crisis communication dynamics?

Key findings

• The majority of detected bots during the Manchester Bombing were classified as benign, primarily sharing news updates and official information.
• Although bot accounts represented a small proportion of total Twitter activity, their tweet volume and engagement levels were disproportionately high.
• Bot activity was concentrated during peak crisis hours, indicating strategic deployment to support real-time information needs.
• No significant evidence of coordinated disinformation campaigns or malicious behavior was observed among the detected bot accounts.
• Bots played a notable role in amplifying credible sources and official statements, contributing to information stability during uncertainty.
• The findings suggest that social bots can serve as valuable tools in crisis communication when used for legitimate, timely information sharing.