Incidental effects of automated retweeting: an exploratory network perspective on bot activity during Sri Lanka’s presidential election in 2015

Chamil Rathnayake, Wayne Buente

Research output: Contribution to journalArticle

Abstract

The role of automated or semiautomated social media accounts, commonly known as “bots,” in social and political processes has gained significant scholarly attention. The current body of research discusses how bots can be designed to achieve specific purposes as well as instances of unexpected negative outcomes of such use. We suggest that the interplay between social media affordances and user practices can result in incidental effects from automated agents. We examined a Twitter network data set with 1,782 nodes and 5,640 edges to demonstrate the engagement and outreach of a retweeting bot called Siripalabot that was popular among Sri Lankan Twitter users. The bot served the simple function of retweeting tweets with hashtags #SriLanka and #lk to its follower network. However, the co-use of #Sri Lanka and/or #lk with #PresPollSL, a hashtag used to discuss politics related to Sri Lanka’s presidential election in 2015, resulted in the bot incidentally amplifying the political voice of less engaged actors. The analysis demonstrated that the bot dominated the network in terms of engagement (out-degree) and the ability to connect distant clusters of actors (betweenness centrality) while more traditional actors, such as the main election candidates and news accounts, indicated more prestige (in-degree) and power (eigenvector centrality). We suggest that the study of automated agents should include designer intentions, the design and behavior of automated agents, user expectations, as well as unintended and incidental effects of interaction.
LanguageEnglish
Pages57-65
Number of pages9
JournalBulletin of Science, Technology and Society
Volume37
Issue number1
Early online date25 Oct 2017
DOIs
Publication statusE-pub ahead of print - 25 Oct 2017

Fingerprint

Sri Lanka
presidential election
twitter
social media
data network
follower
prestige
news
candidacy
election
politics
ability
interaction

Keywords

  • twitter
  • bots
  • Sri Lanka
  • politics

Cite this

@article{9d706d29eb0e4abeaf4cc60db9b82fbb,
title = "Incidental effects of automated retweeting: an exploratory network perspective on bot activity during Sri Lanka’s presidential election in 2015",
abstract = "The role of automated or semiautomated social media accounts, commonly known as “bots,” in social and political processes has gained significant scholarly attention. The current body of research discusses how bots can be designed to achieve specific purposes as well as instances of unexpected negative outcomes of such use. We suggest that the interplay between social media affordances and user practices can result in incidental effects from automated agents. We examined a Twitter network data set with 1,782 nodes and 5,640 edges to demonstrate the engagement and outreach of a retweeting bot called Siripalabot that was popular among Sri Lankan Twitter users. The bot served the simple function of retweeting tweets with hashtags #SriLanka and #lk to its follower network. However, the co-use of #Sri Lanka and/or #lk with #PresPollSL, a hashtag used to discuss politics related to Sri Lanka’s presidential election in 2015, resulted in the bot incidentally amplifying the political voice of less engaged actors. The analysis demonstrated that the bot dominated the network in terms of engagement (out-degree) and the ability to connect distant clusters of actors (betweenness centrality) while more traditional actors, such as the main election candidates and news accounts, indicated more prestige (in-degree) and power (eigenvector centrality). We suggest that the study of automated agents should include designer intentions, the design and behavior of automated agents, user expectations, as well as unintended and incidental effects of interaction.",
keywords = "twitter, bots, Sri Lanka, politics",
author = "Chamil Rathnayake and Wayne Buente",
year = "2017",
month = "10",
day = "25",
doi = "10.1177/0270467617736219",
language = "English",
volume = "37",
pages = "57--65",
journal = "Bulletin of Science, Technology and Society",
issn = "0270-4676",
number = "1",

}

TY - JOUR

T1 - Incidental effects of automated retweeting

T2 - Bulletin of Science, Technology and Society

AU - Rathnayake, Chamil

AU - Buente, Wayne

PY - 2017/10/25

Y1 - 2017/10/25

N2 - The role of automated or semiautomated social media accounts, commonly known as “bots,” in social and political processes has gained significant scholarly attention. The current body of research discusses how bots can be designed to achieve specific purposes as well as instances of unexpected negative outcomes of such use. We suggest that the interplay between social media affordances and user practices can result in incidental effects from automated agents. We examined a Twitter network data set with 1,782 nodes and 5,640 edges to demonstrate the engagement and outreach of a retweeting bot called Siripalabot that was popular among Sri Lankan Twitter users. The bot served the simple function of retweeting tweets with hashtags #SriLanka and #lk to its follower network. However, the co-use of #Sri Lanka and/or #lk with #PresPollSL, a hashtag used to discuss politics related to Sri Lanka’s presidential election in 2015, resulted in the bot incidentally amplifying the political voice of less engaged actors. The analysis demonstrated that the bot dominated the network in terms of engagement (out-degree) and the ability to connect distant clusters of actors (betweenness centrality) while more traditional actors, such as the main election candidates and news accounts, indicated more prestige (in-degree) and power (eigenvector centrality). We suggest that the study of automated agents should include designer intentions, the design and behavior of automated agents, user expectations, as well as unintended and incidental effects of interaction.

AB - The role of automated or semiautomated social media accounts, commonly known as “bots,” in social and political processes has gained significant scholarly attention. The current body of research discusses how bots can be designed to achieve specific purposes as well as instances of unexpected negative outcomes of such use. We suggest that the interplay between social media affordances and user practices can result in incidental effects from automated agents. We examined a Twitter network data set with 1,782 nodes and 5,640 edges to demonstrate the engagement and outreach of a retweeting bot called Siripalabot that was popular among Sri Lankan Twitter users. The bot served the simple function of retweeting tweets with hashtags #SriLanka and #lk to its follower network. However, the co-use of #Sri Lanka and/or #lk with #PresPollSL, a hashtag used to discuss politics related to Sri Lanka’s presidential election in 2015, resulted in the bot incidentally amplifying the political voice of less engaged actors. The analysis demonstrated that the bot dominated the network in terms of engagement (out-degree) and the ability to connect distant clusters of actors (betweenness centrality) while more traditional actors, such as the main election candidates and news accounts, indicated more prestige (in-degree) and power (eigenvector centrality). We suggest that the study of automated agents should include designer intentions, the design and behavior of automated agents, user expectations, as well as unintended and incidental effects of interaction.

KW - twitter

KW - bots

KW - Sri Lanka

KW - politics

U2 - 10.1177/0270467617736219

DO - 10.1177/0270467617736219

M3 - Article

VL - 37

SP - 57

EP - 65

JO - Bulletin of Science, Technology and Society

JF - Bulletin of Science, Technology and Society

SN - 0270-4676

IS - 1

ER -