These Kamala Harris critics do not exist
In which a low-effort spam network floods the replies of popular X accounts with duplicate attacks on the Democratic presidential candidate
The 2024 U.S. presidential election is just over a month away, and it should come as no surprise to anyone that various attempts at influencing the contest via social media manipulation are underway. Over on social media platform X, a network of hundreds of recently-created inauthentic accounts has been bombarding the reply sections of popular posts with repetitively worded attacks on Democratic presidential candidate Kamala Harris. The spammy accounts in question also occasionally amplify various popular right-wing figures, and make use of both stolen and artificially generated images as profile photos.
This reply spam network consists at least 545 X accounts, created between January and September 2024, mostly during the latter portion of that period. Each account has posted no more than 600 times as of September 27th, 2024, and most have only a handful of followers (many of which appear to be fake). The accounts in question were found by searching for replies containing both “Kamala” and “every life”, and then filtering to replies that were duplicated in the dataset at least 10 times in order to weed out false positives.
Almost all of the content posted by this spam network is replies: 88114 of 88364 posts, or 99.7%. Most of these replies are to large accounts that support Kamala Harris, including the official @KamalaHQ X account. These replies are extremely repetitive, with dozens of accounts frequently replying to different posts with the same comment. The network’s recent replies are mostly attacks on Harris, such as “Kamala believes it’s okay to end the life of a baby, but every life is sacred”, which was posted verbatim 59 times by 57 distinct spam accounts. Previous repetitive replies include pro-Trump and anti-Biden comments, mixed with apolitical sports-related chatter.
The activity of this spam network ramped up over time as additional accounts came into service, with a significant increase in volume in mid-June. Content posted at the time of this increase is largely focused on attacking Joe Biden, who was at that point still the presumptive Democratic nominee for the U.S. presidency. Some of the network’s posts also target Senate and House races, expressing spammy support for Republican candidates and criticizing Democrats.
In addition to the reply spam, the accounts in the network also repost content, albeit infrequently. The material reposted by the network is mostly from major right-wing influencers, with large meme pages thrown into the mix as well. The five accounts most frequently amplified are Elon Musk (@elonmusk), Libs of Tiktok (@libsoftiktok), Amazing Nature (@AMAZINGNATURE), Charlie Kirk (@charliekirk11), and Cat Turd (@catturd2).
The accounts in the network largely use drawings, cartoons, and logos as their profile images, but there are a few exceptions that serve to underscore the inauthenticity of the whole effort. A handful of the accounts use AI-generated faces as avatars, specifically StyleGAN-generated face images similar to those served by thispersondoesnotexist.com.
Some of the spam accounts use genuine photos of real people as avatars. Predictably, these photographs are plagiarized, usually from random people’s social media feeds or the staff directories of random business websites. (Many of of the stolen photos include children; those reverse image searches are therefore excluded from the collage below.) The stolen images were easily identified using Google reverse image search; TinEye was strangely ineffective here.
The Russian Chinese network continues apace, yet this time people are more informed about them. Thank you for such great work.
We're all so happy that Leon Musk cleaned up Twitter! Who needs employees when software does such an amazing job of content moderation??