How does one find astroturf networks? One method is to choose a word or phrase, plot a histogram of the creation dates of accounts tweeting that word/phrase, and look for spikes. Several such spikes show up among accounts with recent tweets containing "casino".
If the spikes in creation dates are the result of someone creating making en masse, batches of similar-looking accounts will show up when one looks at the accounts created during the spikes. Here are some examples from the casino dataset, with possible batches highlighted in red.
The late February and early March batches from the casino dataset all use the same naming scheme, tweet using the same app ("Twitter Web App"), retweet many of the same tweets, and tweet identical tweets on multiple accounts, so these accounts are likely a single operation.
Sometimes the initial search will only uncover part of a bot/sock astroturf network. One can frequently find more accounts that are part of a network by looking at other accounts retweeting the same tweets, and exploring the followers/followees of the initial accounts.
Following this procedure with the late February/early March batch-created accounts yields 516 accounts that appear to be part of the network, all with names that appear to be randomly selected pairs of English words.
What do these accounts actually do? They retweet and reply to cryptocurrency accounts, mostly. @Roobet is the most frequent beneficiary of their astroturfing - several of its tweets have been retweeted, replied to, and liked by all 516 accounts in the network.
These accounts also follow each other profusely, in two separate groups. One cluster consists solely of accounts with late February 2021 creation dates, while the other is mostly March 2021 accounts with a few February 2021 accounts thrown in.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
As Bluesky approaches 30 million users, people who run spam-for-hire operations are taking note. Here's a look at a network of fake Bluesky accounts associated with a spam operation that provides fake followers for multiple platforms.
cc: @ZellaQuixote
This fake follower network consists of 8070 Bluesky accounts created between Nov 30 and Dec 30, 2024. None has posted, although some have reposted here and there. Almost all of their biographies are in Portuguese, with the exception of a few whose biographies only contain emoji.
The accounts in this fake follower network use a variety of repeated or otherwise formulaic biographies, some of which are repeated dozens or hundred of times. Some of the biographies begin with unnecessary leading commas, and a few consist entirely of punctuation.
It's presently unclear why, but over the past year someone has created a network of fake Facebook accounts pretending to be employees of the Los Angeles Dodgers. Many of the accounts in this network have GAN-generated faces.
cc: @ZellaQuixote
This network consists of (at least) 80 Facebook accounts, 48 of which use StyleGAN-generated faces as profile images. The remaining 32 all use the same image, a real photograph of a random person sitting in an office.
As is the case with all unmodified StyleGAN-generated faces, the main facial features (especially the eyes) are in the same position on all 48 AI-generated faces used by the network. This anomaly becomes obvious when the faces are blended together.
None of these chefs exist, as they're all AI-generated images. This hasn't stopped them from racking up lots of engagement on Facebook by posting AI-generated images of food (and occasional thoughts and prayers), however.
cc: @ZellaQuixote
These "chefs" are part of a network of 18 Facebook pages with names like "Cook Fastly" and "Emily Recipes" that continually post AI-generated images of food. While many of these pages claim to be US-based, they are have admins in Morocco per Facebook's Page Transparency feature.
Between them, these 18 Facebook "chef" pages have posted AI-generated images of food at least 36,000 times in the last five months. Not all of the images are unique; many have been posted repeatedly, sometimes by more than one of the alleged chefs.
Can simple text generation bots keep sophisticated LLM chatbots like ChatGPT engaged indefinitely? The answer is yes, which has some potentially interesting implications for distinguishing between conversational chatbots and humans.
For this experiment, four simple chatbots were created:
• a bot that asks the same question over and over
• a bot that replies with random fragments of a work of fiction
• a bot that asks randomly generated questions
• a bot that repeatedly asks "what do you mean by <X>?"
The output of these chatbots was used as input to an LLM chatbot based on the 8B version of the Llama 3.1 model. Three of the four bots were successful at engaging the LLM chatbot in a 1000-message exchange; the only one that failed was the repetitive question bot.
The spammers behind the "Barndominium Gallery" Facebook page have branched out into AI-generated video and started a YouTube channel with the catchy name "AY CUSTOM HOME". The results are just about as craptastic as you'd expect.
In this synthetically generated aerial video of a (nonexistent) barndominium under construction, the geometry of the roof changes, a blue building appears, and a tree vanishes, all in the course of just three seconds.
This AI-generated barndominium features a long AI-generated porch with some chairs on it. Exactly how many chairs there are depends on what angle you look at it from, however, as the chair on the left splits into three chairs as the camera pans.
Some observations regarding @Botted_Likes (permanent ID 1459592225952649221)...
First, "viral posts which don't result in follower growth and have very little engagement in the reply section" is not a useful heuristic for detecting botted likes. Why not?
cc: @ZellaQuixote
"Viral posts that do not result in follower growth" is not a valid test for botting, because posts from large accounts often go viral among the large account's existing followers but do not reach other audiences, resulting in high like/repost counts but little/no follower growth.
"Very little engagement in the reply section" doesn't work for multiple reasons (some topics spur debate and some don't, some people restrict replies, etc)
Hilariously, @Botted_Likes seems to be ignoring their own criteria, as many of the posts they feature have tons of replies.