I didn’t intend to start this project. It happened gradually over the past year as I read more about the tactics of the Russian-based internet Research Agency (IRA), which was a major part of the early indictments issued by the Justice Department in cooperation with Special Prosecutor, Robert Mueller III. While Trump supporters crow, in person and on social media, that “no Russians were in the voting booth,” the presence of QAnon at rallies (seriously; when are we going to have serious campaign reform that bans presidential “rallies” such as these?) is evidence enough that the disinformation campaigns spearheaded by Russian hackers have been internalized as fact by an overwhelming majority of the self-described deplorables. There is nothing quite like the arrogance of false certainty.
I’ve been dealing with some health issues and I am incapable of not working on multiple things at once, so I began spending more time on Twitter. This was made possible largely by the staggeringly tumultuous past two weeks. Trump has had multiple tweet storms that descended into tantrums. I began to notice patterns in the more outrageous comments. Soon, I was discovering entire conversations with bots or highly-moderated accounts. This has heated up exponentially since a contingent of pro-Trump Black pastors visited the White House. African-American leaders from around the country have taken to Twitter and Facebook to decry not only the behavior of particular pastor but also the Administration’s woeful record on important issues around race. This “intraracial” conversation is the perfect opportunity for Russian bots to fulfill their mission to further divides and spread disinformation. Surprise, surprise. Today, Trump tweeted a Rasmussen poll citing a rising level of support among Blacks. As expected, the bots are most active.
What kind of numbers are we talking about?
Mine was in no way a scientific study. I am finishing a doctorate and I know what is required for a quantitative study. This ain’t it. However, I hope this is useful for the average person who wants to be able to use Twitter to connect with real people who are interested in more than shitpot stirring.
Had I thought through this before I did it, I would have taken screenshots of the conversations. I wasn’t intending on writing this but I was just curious about how many I would find in a day, which grew into a weekend, which grew into a week, which grew into two weeks. Each time I would find a confirmed bot, I would report and then block. What you can see below are screenshots of about 95% of the accounts I identified.
I am going to place the confirmed number of bots at 110. I absolutely could have found three times that amount if I wanted to dedicate myself to the task 40 hours/week. I absolutely didn’t and don’t want to do that, but I conservatively say that I dedicated about 30 hours over the course of two weeks.
Okay, 110 out of how many?
I definitely got better with this over time (see below), so there were a number of false checks early on that I didn’t even bother with after a week of steady investigation. I estimate that I checked around 300-350 accounts. So, we’re talking about 1/3. Again, this is not scientific and I am not saying that a full 1/3 of accounts commenting on Trump’s post are bots; I actually think it is higher, but I leave it to those much more qualified than am I to do proper studies.
Why does this even matter?
I think the saddest thing for me was the 200 or so accounts I checked that were not bots but rather belonged to real people parroting the rhetoric of the fake accounts. #walkaway is a perfect example. This is a manufactured “movement,” replete with stock photographs of Black people superimposed with fake quotes about how they are leaving the Democratic party. Almost every instance of a legitimate person using #walkaway, that person was White and they followed the fake accounts of supposedly Black, Republican Americans. It is like the Twitter version of “I can’t be racist because I have Black friends.
Even more insane are the conversations between highly moderated accounts pretending to be Black. One, between @sugarthegirl and @spoilrottenpup, was about how they can no longer talk to their families because they support Trump. Again, I acknowledge I should have captured the conversation itself, but I have gotten so used to automatically reporting and blocking that I went too quickly. Alas, you can see the profile pages.
Sugar the Girl did, it should be mentioned, send me a snapshot of a handwritten note with the date and a message telling me to suck it. I would have been more impressed with a photo of a real person. It does show, however, that there are humans behind these accounts, but it is highly doubtful that they are who they claim to be.
How to identify Bots
Some of what I learned might be well-known to others who are more informed in the field. I’m just a dude with a blog who noticed some stuff. The first thing you need is botcheck.me . To save yourself time, here are a few things that can help zero in on suspicious accounts.
- A string of numbers by the name. If you see a name like DanEboy12800473, that’s a good candidate. This does not always hold, as it appears some people purposefully have such usernames so they can goad others into accusing them of being bots, only to respond with something like, “Typical libtard. Everyone who doesn’t agree with you is either a Russian or a bot #walkaway”
- They are relatively new to Twitter. Again, this is in no way a guarantee, but it has been evident in enough cases that it is worth a mention.
- If they mention having been “shadow banned” or have a “new account” because they’re being persecuted as conservatives.
- Their self-description is just a little too filled with accomplishments that are not backed up with any professional tweets, the page has links to nothing other than pro-Trump propaganda, and they use lots of hashtags associated with Russian meddling.
- Identities that seem incongruous, like “Muslims for the travel ban” or “Mexicans for the wall.” Sadly, this also isn’t a guarantee. In all cases, use botcheck.me.
- If they swoop in, post something provocative, and then move on. This is a tactic I saw a lot. It stirs the pot, and if enough real people get involved, the bot doesn’t comment much because someone else has taken on the role of provocateur.
What to do?
I don’t recommend spending as much time as I have on chasing down bots. It has certainly impacted my mental health and I find that even though I know the accounts are “fake,” the impact is not. What the intelligence agencies have been saying about Russia’s ongoing efforts is playing out right in front of our eyes. Each time OfVald has another rally and we see the rabid, ignorant, hate-filled supporters screaming propaganda while, without any trace of irony, declaring their superiority to the dumb “libs,” I feel both sad and angry. But this is part of the effort. It is meant to wear us down. It is meant to make us fear that Trump’s actual supporters outnumber us.
Twitter can be great, but it takes diligence if you want to have any sort of meaningful conversations or exchanges of ideas. As for me, I am going to take a break from reading the comment threads and focus on thoughtful pieces longer than 280 characters.