If you've been wanting to meet an automated Reagan conservative who worries about our future by spamming Twitter with links to poorly-source articles about "voter fraud" on a website you've never heard of, @keichri is the account for you. #FridayFeeling
Almost all (98.5%) of @keichri's recent tweets are automated via dlvr(dot)it. This account has apparently dabbled in automation before, although at least two of the apps it was using have subsequently been shut down (erased7812941 and erased13330229).
The primary mission of @keichri appears to be to promote various right-wing websites, usually obscure ones. Almost all of its tweets since March 2020 link to speakingaboutnews(dot)com, with a previous three and a half year binge of spamming links to conservativeread(dot)com.
Many of @keichri's tweets contain hashtags as well as links to websites, with #TCOT being the most common. In recent days, it has focused on 2020 election hashtags, including conspiracy theory-related tags such as #HunterBiden, #BidenCorruption, #ElectionFraud, and #VoterFraud.
Although most of @keichri's followers look like random #MAGA accounts, there are some interesting anomalies in the 2013-2014 era where it appears to have been followed by various batch-created accounts with zero likes.
311 of @keichri's batch-created followers are (mostly dormant) automated accounts, mostly Russian-language. These accounts have tweeted via a whopping 149 different apps. We plan to further explore of this network/these networks and will post if we find anything interesting.
(further exploration of the bot network following @keichri)
We found 25574 automated accounts that we believe to be part of this botnet. All were created in batches between 2012 and 2014, and they use repetitive naming schemes. (This is probably not the whole network, we stopped searching when the pace of finding more accounts slowed.)
These accounts are mostly dormant, but back when they were active they tweeted via a wide variety of automation apps (883 apps total). Almost all of their content is in Russian.
It's a Thursday, which is an excellent time to look at a recently created fake follower/retweet botnet that uses GAN-generated profile pics (similar to those produced by thispersondoesnotexist.com). #ThursdayShenaniGANs
This botnet consists of 164 accounts created between October 9th and October 13th, 2020. All of them either have never tweeted or post all of their tweets (nearly all of which are retweets) via "Twitter Web App".
Here are the profile pics of all 164 accounts in this botnet, as well as the result of blending them together. Like all unmodified pics produced by StyleGAN (the neural network used by thispersondoesnotexist.com), the eyes, nose, and mouth are in the same location on every image.
It's election day in the USA, and like any other day, porn/adult services botnets are spewing their low-effort salacious spam. Here's a Korean-language network we recently stumbled upon. #TuesdayThoughts#GOTV2020
This pornbot network consists of 162 accounts created between October 14th and November 3rd, 2020. All have default profile pics, randomly generated 15 character usernames, and send almost all of their tweets via automation app cubi(dot)so.
The bots in this network tweet pretty much 24/7, with occasional gaps at what appear to be random times. The entire botnet was inactive for most of October 30th and 31st, and many of the accounts have few or no tweets prior to those dates.
We quite frankly have no idea why these accounts are all tweeting the word "momotarousuper1" (or what momotarousuper1 even means), but let's take a look at them anyway. #TuesdayThoughts
We found 118 accounts tweeting "momotarousuper1" over and over, all created in December 2012, March 2013, or April 2013. Their names are quite repetitive, with most account names beginning with "momotalou" and ending with X's and numbers.
All 118 accounts in this botnet send all of their tweets via automation service twittbot(dot)net, and they tweet on very similar schedules. What do they tweet?
How did @SVNewsAlerts, a popular purveyor of melodramatic "breaking news" tweets that frequently get retweeted far and wide, rack up 74.1K followers in 6 months? #SundayThoughts
The first accounts to follow @SVNewsAlerts look like #MAGA followback accounts (similar numbers of followers/following). @SVNewsAlerts only follows one account, though (its own alt, @SVNews_Reporter), so surely it didn't build its massive audience via followbacks. . . or did it?
As it turns out, @SVNewsAlerts (permanent ID 1263205289618214913) wasn't always named @SVNewsAlerts. For the first couple months of its Twitter existence, it was named @republican_owl, and at least one tweet indicating that it used to follow back is still live.
We've done a decent amount of research on the use of GAN-generated images over the last two years, mostly fake face "photos" such as those produced by thispersondoesnotexist(dot)com. Here are all of our related threads in one place.
GAN is an abbreviation for "generative adversarial network", the AI technique used to produce these images (and other things). Here's a "brief" description of how StyleGAN, the GAN behind thispersondoesnotexist(dot)com, works. machinelearningmastery.com/introduction-t…
Here's the first network we found in the wild using GAN-generated face pics (29 of 52 accounts, created in two batches). The profiles featured repetitive biographies mentioning family, friends, country, and weapons.