Bot comments Report July 2024
The dataset of toxic bot comments detected in July 2024, spanning from July 1 to July 31, included over 550 English messages (n=558). The average toxicity score of the bot content was relatively low, at 0.20. However, a notable spike occurred on July 9, 2024, potentially linked to the civil war in Myanmar, where the liberation army claimed that shelling by the national army killed seven refugees (see the full list of events here).
Despite the low overall toxicity, several messages featured highly problematic content with toxic keywords such as evil Jews, Jews, moron, genocidal, piece of shit, deep state, treasonous, the white man, disgusting, and ugly.
Toxicity frequently involved:
Politics (10%), Racism, Ridicule, and/or Religion.
Most frequent combination: Politics + Racism.
Threatening and Disinformation content: Both categories made up 3% of the messages (see explanations of the categories here).
The word cloud below highlights the 50 most frequently used toxic keywords from July 2024. The size of each word reflects its frequency relative to others, with trump and ho being the most commonly used, followed by black, racist, and Indians.
These words form the top 50 of keywords used by bot comments in July 2024 (THREAT = bold)
The table below highlights the various toxic categories detected in the July 2024 dataset, along with their respective frequencies (represented by black dots and explained further below). No single category was significantly more prominent than the others. However, content targeting Jews had the highest toxicity score of 0.75 (on a scale from 0 to 1, where 1 represents the highest level of toxicity), indicating a particularly severe level of harmful content within that category.
The bar chart below reveals that the most prominent categories with the highest toxicity scores in July 2024’s dataset of toxic bot comments were Ridicule, Profanity, and Politics. However, the majority of the bot content fell under the Politics category, indicating that the primary objective of the toxic bot comments was to influence political discussions.
Number of toxic messages per category
The keyword network below visualizes the main focus of toxic bot comments detected in July 2024. Different colors represent clusters of words that frequently appeared together, highlighting key narratives promoted by the bot content. There is a clear focus on American politics, with prominent keywords such as Kamala, Biden, Joe, Vivek, and Trump frequently occurring together. This indicates that a significant portion of the toxic bot activity was centered around discussions involving U.S. political figures and events.
