Bot comments Report May 2025
The dataset of toxic bot comments detected in May 2025, spanning from May 1 to May 23, included 362 English messages. The average toxicity score is 0.15, which is quite low. However, a significant peak in toxic activity was observed on May 18, 2025 (see this list of events that could have influenced it).
Threatening content: about 15% of the messages was categorized as threatening
Disinformation content: about 1% of the messages was categorized as disinformation
The most frequent combination of categories: Politics + Religion
The word cloud below highlights the 50 most frequently used toxic keywords in February 2025. The size of each word reflects its frequency relative to other terms. Notably, fire was the most frequently used keyword, followed by israel and genocide.
The table below highlights the different toxic categories detected within the May 2025 dataset, along with their respective frequencies (represented by black dots and further explained below). In this month, the categories Poiticis, Religion and Threat. The category Threat recieved the highest toxicity score, 0.25, followed by Poitics, Racism and Religion with 0.15.
The barchart below shows that most of the bot content fell withint he categories Poitics, Religion and Threat. Those are also the categories that involve the most toxic content, with a score of > 0.2 (see the explanation of the different categories here).
The keyword network below illustrates the main focus of toxic bot comments detected in March 2025. Different colors represent clusters of words frequently used together, revealing the bot’s primary narratives. In this case, the keywords european and jews appear often together and the words israel, pray, fire and genocide often appeared together.