Cyborgs for strategic communication on social media

TL;DR

This study defines Cyborgs on social media as accounts that alternately appear as bots and humans over time, using a longitudinal, data-driven approach on two large Twitter datasets (Coronavirus and 2020 US Elections). It establishes quantitative thresholds for Cyborg identification: more than $3$ bot/human flips and an average bot-probability change of at least $0.10$ during flips, grounded by BotHunter outputs with a $0.70$ bot-class threshold. Through network analysis, stance propagation, and topic modeling, the paper shows Cyborgs occupy central positions in the interaction network, engage in both sides of debates, and are predominantly used for strategic communications by Activists and Renowned Personalities, with substantial longevity and notable but not universal suspension patterns. The work contributes a repeatable framework for Cyborg detection and a nuanced characterization of their communicative role, suggesting they function as durable StratCom instruments capable of broad reach and cross-cutting influence in online discourse.

Abstract

Social media platforms are a key ground of information consumption and dissemination. Key figures like politicians, celebrities and activists have leveraged on its wide user base for strategic communication. Strategic communications, or StratCom, is the deliberate act of information creation and distribution. Its techniques are used by these key figures for establishing their brand and amplifying their messages. Automated scripts are used on top of personal touches to quickly and effectively perform these tasks. The combination of automation and manual online posting creates a Cyborg social media profile, which is a hybrid between bot and human. In this study, we establish a quantitative definition for a Cyborg account, which is an account that are detected as bots in one time window, and identified as humans in another. This definition makes use of frequent changes of bot classification labels and large differences in bot likelihood scores to identify Cyborgs. We perform a large-scale analysis across over 3.1 million users from Twitter collected from two key events, the 2020 Coronavirus pandemic and 2020 US Elections. We extract Cyborgs from two datasets and employ tools from network science, natural language processing and manual annotation to characterize Cyborg accounts. Our analyses identify Cyborg accounts are mostly constructed for strategic communication uses, have a strong duality in their bot/human classification and are tactically positioned in the social media network, aiding these accounts to promote their desired content. Cyborgs are also discovered to have long online lives, indicating their ability to evade bot detectors, or the graciousness of platforms to allow their operations.

Figures (3)

• Figure 1: Proportion of number of agents against number of flips within a month. Cyborgs are agents that flip more than thrice, which are the 75th-percentile of agents.
• Figure 2: Proportion of the number of agents against the absolute difference of bot probability score when the bot classification flips. Cyborgs are agents that flip bot classification with more than 0.10 score differences.
• Figure 3: Standard deviations of bot scores per type of agent.