Suarez-Lledo Victor, Ortega-Martin Esther, Carretero-Bravo Jesus, Ramos-Fiol Begoña, Alvarez-Galvez Javier
Computational Social Science DataLab, University Institute of Research for Sustainable Social Development (INDESS), University of Cadiz, Jerez de la Frontera, Spain.
Department of Sociology, University of Granada, Granada, Spain.
JMIR Infodemiology. 2025 Jan 9;5:e50021. doi: 10.2196/50021.
During the COVID-19 pandemic, social media platforms have been a venue for the exchange of messages, including those related to fake news. There are also accounts programmed to disseminate and amplify specific messages, which can affect individual decision-making and present new challenges for public health.
This study aimed to analyze how social bots use hashtags compared to human users on topics related to misinformation during the outbreak of the COVID-19 pandemic.
We selected posts on specific topics related to infodemics such as vaccines, hydroxychloroquine, military, conspiracy, laboratory, Bill Gates, 5G, and UV. We built a network based on the co-occurrence of hashtags and classified the posts based on their source. Using network analysis and community detection algorithms, we identified hashtags that tend to appear together in messages. For each topic, we extracted the most relevant subtopic communities, which are groups of interconnected hashtags.
The distribution of bots and nonbots in each of these communities was uneven, with some sets of hashtags being more common among accounts classified as bots or nonbots. Hashtags related to the Trump and QAnon social movements were common among bots, and specific hashtags with anti-Asian sentiments were also identified. In the subcommunities most populated by bots in the case of vaccines, the group of hashtags including #billgates, #pandemic, and #china was among the most common.
The use of certain hashtags varies depending on the source, and some hashtags are used for different purposes. Understanding these patterns may help address the spread of health misinformation on social media networks.
在新冠疫情期间，社交媒体平台一直是信息交流的场所，包括与假新闻相关的信息。也有一些程序设定的账号来传播和放大特定信息，这可能会影响个人决策，并给公共卫生带来新挑战。
本研究旨在分析在新冠疫情爆发期间，与人类用户相比，社交机器人在与错误信息相关的话题上如何使用主题标签。
我们选择了与信息疫情相关的特定主题的帖子，如疫苗、羟氯喹、军事、阴谋、实验室、比尔·盖茨、5G和紫外线。我们基于主题标签的共现构建了一个网络，并根据帖子的来源对其进行分类。使用网络分析和社区检测算法，我们识别出在信息中倾向于一起出现的主题标签。对于每个主题，我们提取了最相关的子主题社区，即相互关联的主题标签组。
这些社区中机器人账号和非机器人账号的分布不均衡，某些主题标签集在被分类为机器人或非机器人的账号中更为常见。与特朗普和匿名者Q运动相关的主题标签在机器人账号中很常见，还识别出了带有反亚裔情绪的特定主题标签。在疫苗相关的机器人账号最多的子社区中，包括#比尔盖茨#、#大流行#和#中国#的主题标签组是最常见的之一。
某些主题标签的使用因来源而异，一些主题标签被用于不同目的。了解这些模式可能有助于应对社交媒体网络上健康错误信息的传播。
