Twitter bots are a controversial element of the platform, and their negative impact is well known. In the field of scientific communication, they have been perceived in a more positive light, and the accounts that serve as feeds alerting about scientific publications are quite common. However, despite being aware of the presence of bots in the dissemination of science, no large-scale estimations have been made nor has it been evaluated if they can truly interfere with altmetrics. Analyzing a dataset of 3,744,231 papers published between 2017 and 2021 and their associated 51,230,936 Twitter mentions, our goal was to determine the volume of publications mentioned by bots and whether they skew altmetrics indicators. Using the BotometerLite API, we categorized Twitter accounts based on their likelihood of being bots. The results showed that 11,073 accounts (0.23% of total users) exhibited automated behavior, contributing to 4.72% of all mentions. A significant bias was observed in the activity of bots. Their presence was particularly pronounced in disciplines such as Mathematics, Physics, and Space Sciences, with some specialties even exceeding 70% of the tweets. However, these are extreme cases, and the impact of this activity on altmetrics varies by speciality, with minimal influence in Arts & Humanities and Social Sciences. This research emphasizes the importance of distinguishing between specialties and disciplines when using Twitter as an altmetric.

翻译：Twitter机器人是该平台上一个颇具争议的元素，其负面影响广为人知。在科学传播领域，这些机器人则被赋予了较为积极的看法，那些用于推送科学出版物动态的账户相当常见。然而，尽管人们意识到机器人存在于科学传播中，但尚未开展大规模评估，也未检验它们是否真正干扰了替代计量指标。通过分析2017年至2021年间发表的3,744,231篇论文及其相关的51,230,936条Twitter提及数据，我们的目标是确定被机器人提及的论文数量，以及这些机器人是否会扭曲替代计量指标。利用BotometerLite API，我们根据Twitter账户具有机器行为的可能性对其进行了分类。结果显示，共有11,073个账户（占用户总数的0.23%）表现出自动化行为，贡献了所有提及量的4.72%。在机器人的活动模式中观察到显著偏差。它们的出现在数学、物理学和空间科学等学科中尤为突出，某些专业领域的提及中甚至超过70%来自机器人。然而，这些属于极端案例，此类活动对替代计量指标的影响因专业而异，在艺术与人文学科以及社会科学中影响极小。本研究强调了在将Twitter用作替代计量工具时，区分不同专业与学科的重要性。