Social Cyber Geographical Worldwide Inventory of Bots
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910807085809664 |
|---|---|
| author | Ng, Lynnette Hui Xian Carley, Kathleen M. |
| author_facet | Ng, Lynnette Hui Xian Carley, Kathleen M. |
| contents | Social Cyber Geography is the space in the digital cyber realm that is produced through social relations. Communication in the social media ecosystem happens not only because of human interactions, but is also fueled by algorithmically controlled bot agents. Most studies have not looked at the social cyber geography of bots because they focus on bot activity within a single country. Since creating a bot uses universal programming technology, bots, how prevalent are these bots throughout the world? To quantify bot activity worldwide, we perform a multilingual and geospatial analysis on a large dataset of social data collected from X during the Coronavirus pandemic in 2021. This pandemic affected most of the world, and thus is a common topic of discussion. Our dataset consists of ~100 mil posts generated by ~31mil users. Most bot studies focus only on English-speaking countries, because most bot detection algorithms are built for the English language. However, only 47\% of the bots write in the English language. To accommodate multiple languages in our bot detection algorithm, we built Multilingual BotBuster, a multi-language bot detection algorithm to identify the bots in this diverse dataset. We also create a Geographical Location Identifier to swiftly identify the countries a user affiliates with in his description. Our results show that bots can appear to move from one country to another, but the language they write in remains relatively constant. Bots distribute narratives on distinct topics related to their self-declared country affiliation. Finally, despite the diverse distribution of bot locations around the world, the proportion of bots per country is about 20%. Our work stresses the importance of a united analysis of the cyber and physical realms, where we combine both spheres to inventorize the language and location of social media bots and understand communication strategies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_18839 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Social Cyber Geographical Worldwide Inventory of Bots Ng, Lynnette Hui Xian Carley, Kathleen M. Social and Information Networks Social Cyber Geography is the space in the digital cyber realm that is produced through social relations. Communication in the social media ecosystem happens not only because of human interactions, but is also fueled by algorithmically controlled bot agents. Most studies have not looked at the social cyber geography of bots because they focus on bot activity within a single country. Since creating a bot uses universal programming technology, bots, how prevalent are these bots throughout the world? To quantify bot activity worldwide, we perform a multilingual and geospatial analysis on a large dataset of social data collected from X during the Coronavirus pandemic in 2021. This pandemic affected most of the world, and thus is a common topic of discussion. Our dataset consists of ~100 mil posts generated by ~31mil users. Most bot studies focus only on English-speaking countries, because most bot detection algorithms are built for the English language. However, only 47\% of the bots write in the English language. To accommodate multiple languages in our bot detection algorithm, we built Multilingual BotBuster, a multi-language bot detection algorithm to identify the bots in this diverse dataset. We also create a Geographical Location Identifier to swiftly identify the countries a user affiliates with in his description. Our results show that bots can appear to move from one country to another, but the language they write in remains relatively constant. Bots distribute narratives on distinct topics related to their self-declared country affiliation. Finally, despite the diverse distribution of bot locations around the world, the proportion of bots per country is about 20%. Our work stresses the importance of a united analysis of the cyber and physical realms, where we combine both spheres to inventorize the language and location of social media bots and understand communication strategies. |
| title | Social Cyber Geographical Worldwide Inventory of Bots |
| topic | Social and Information Networks |
| url | https://arxiv.org/abs/2501.18839 |