Control of Information by a Few: Incoordinated behavior of social bots in information dissemination

Social media platforms facilitate the spread of information through posts and shares, with an increasing influence from automated accounts. As AI technologies complicate the monitoring of individual accounts, it becomes crucial to address the unrealistic coordination behaviors exhibited by social bots. This study examines account incoordination in information dissemination by analyzing characteristics, network structures, and dynamic patterns using a co-occurrence network approach. We analyze a dataset of 3,823,020 tweets related to the Bucha event, spanning 959,468 accounts, and extract the interaction network of a critical minority. Accounts are categorized into three types based on their dissemination patterns: government or media accounts, social bots, and human users. Our findings reveal that media or government accounts are the primary sources of information, with both social bots and humans amplifying their messages. Unlike humans, social bots rarely cite other bots as sources, which is a key distinction. Social bots play a significant role in accelerating the spread of media messages and, in some cases, manipulating information flow. These findings highlight the need to monitor and regulate social bot activities, particularly in relation to media and government sources, to maintain the integrity of public discourse.