Analysis Q3 2025 toxic bot comments
The dataset of toxic bot comments detected between July 01, 2025 and September 24, 2025, has over 17.5K English messages (n=19,097). The average toxicity score was 0.20, which is quite low. However, there was a notable peak (2 × SD) on July 25, 2025. Several key events could have influenced the spike in toxic bot comments activity but here is a full list of possible events. Despite the low overall toxicity score, the dataset contained problematic and harmful content with the toxic keywords slayer, demon, rape, propaganda, jewish donors, flesh, final solution, deep state, zio, and murder. “Neo-nazi's and Zio-Nazi's are the reason this world is in such bad shape. Them and their cronies start wars, steal resources, rape, torture and murder innocent civilians. Steal their resources and the[...]” (source: unknown).
Prominent toxic categories:
Religion (10%)
Politics
Racism
Ridicule
Threatening content: about 5% of messages fall within this category.
Most frequent combination: Politics + Religion.
About 125 messages (1%) were very toxic, with a score > 0.8.
Below is a visualization of the top 50 toxic keywords from the second quarter of 2025. The size of the words reflects their frequency relative to others. Notably, fira and allah were the most frequently used keywords, followed by rmassacre, carter, oppressed, etc..
The table below highlights the various toxic categories detected during Q3 of 2025 and their respective frequencies (represented by black dots and explained further below). The most prominent categories were Ridicule, Racism, Sexism, Politics, Religion and Threat. The categoy Jews involved the bot comments content with the highest toxicity score, 0.75 (on a scale from 0 to 1, where 1 indicates the highest level of toxicity). The keywords show the focus of most of the toxic bot comments.
The bar chart below highlights that the most prominent category within the toxic bot content during Q3 of 2025 was Politics, followed by Religion. Ridicule, Racism, Politics, Relgion, Threat and Untruth all contained bot comments with a toxicity score > 0.8 (see the explanation of the different categories here).
The keyword network below visualizes the main focus of toxic bot content detected during Q1 of 2025. Different colors represent clusters of words frequently used together, highlighting key narratives. Interestingly, the main narrative in this visualization show a focus on American singer Chapelle Roan.