Spot the Troll
September 27, 2020 12:20 PM   Subscribe

The Clemson University Media Forensics Hub presents: Spot the Troll: A quiz and analysis of social media trolls originating from Russia. Train your filter.

Caveat: their definition of "troll" is specific to mass media campaigns with fake users. Some may consider the non-troll accounts trolls by the original definition, which permits trolling by real live channers humans.
posted by jjray (33 comments total) 20 users marked this as a favorite
 
I'm no good at this, as I just assume all "social media" is fake, malicious, bad-faith, propaganda, and/or cult-fodder.
posted by glonous keming at 12:58 PM on September 27, 2020 [6 favorites]


Bernie Sanders Got A Perfect Score On This Test, Joe Biden Wouldn't Even Take It! #BLM #true #wisconsinforward
posted by Riki tiki at 12:59 PM on September 27, 2020 [5 favorites]


I just assume all "social media" is fake, malicious, bad-faith, propaganda, and/or cult-fodder.

Well, it's not, but you do you.
posted by Ahmad Khani at 1:02 PM on September 27, 2020 [1 favorite]


This is valuable work, although the people who need it most will never see it.

Not that they’d really care where the hate memes come from anyway if they agree with them so hard, so maybe it’s pointless?

I enjoyed it, though.
posted by dragstroke at 1:12 PM on September 27, 2020 [1 favorite]


Well, it's not, but you do you.

#notallsocialmedia
posted by Johnny Assay at 1:13 PM on September 27, 2020 [1 favorite]


I am good at this in real life and was super bad at this quiz. Mostly because I assume more people are trolls than maybe really are? And when I took it I even knew one of the people personally! I would have enjoyed this if there were more randomized examples so you could try again.

Maybe we could not turn this thread into a total referendum on social media right out of the gate? Troll-spotting and learning about Russian disinfo is one particular thing which is what this is about.
posted by jessamyn at 1:15 PM on September 27, 2020 [15 favorites]


Yeah, I came here to post because I was surprised how hard it was - hence my caveat about their definition of trolls vis a vis, well, my own I guess.
posted by jjray at 1:19 PM on September 27, 2020 [3 favorites]


I was better at this than I expected to be. I missed one because I wasn't sure whether it was a fake account promoting a scam website or a real account promoting a scam website.

But, honestly, even if most people were really good at spotting fake accounts, I don't know how much it would help. Most disinformation of this type is meant to be shared virally, and most people aren't reblogging from profile pages, but from timelines.

Half of the problem is that people reblog without thinking or checking at all.
posted by Kutsuwamushi at 1:31 PM on September 27, 2020 [4 favorites]


I got 6 out of 8. Funnily enough, I assumed the feminist troll was real, which is bad, because I'm probably the target audience for fake-feminist trolls. And I guess I should have known that the chemtrails person was real, not a troll, but I'm actually not sure that I distinguish between trolls (in the sense they mean) and professionally-unhinged people. I'm not going to engage with that account either way.
posted by ArbitraryAndCapricious at 1:34 PM on September 27, 2020 [7 favorites]


Part of the problem is that real people retweet posts from trolls. So you can get IRA troll posts from Russia via real people. So you get Trolls and Bots and then the genuine human's who have been parasitized by Trolls and Bots.
posted by srboisvert at 1:35 PM on September 27, 2020 [6 favorites]


[squints, looks suspiciously at thread]
posted by clawsoon at 1:42 PM on September 27, 2020 [4 favorites]


I pretty much noped out of the article when #3 was described as a "caring" person when they were sharing such hateful filth.

I don't much need to know the difference between a foreign fake account posting in bad faith, and a "Real American" posting in bad faith.
posted by explosion at 2:50 PM on September 27, 2020 [6 favorites]


Oof. 4/8, but yeah, chemtrails guy. I sort of question the implication that all trolls are foreign agents. There are more than enough people out there trolling and shitposting because their stunted emotional state leads them to believe that it’s a fun thing to do. It’s just that some folks realized they could make money off of the shittiness, and then others realized, hey, have you seen the shit made up by this emotionally stunted 14 year old? And people believe it? Holy shit, let’s fuck with global elections.

Chemtrails guy is a troll. It’s a big umbrella, covering a lot of assholes.
posted by Ghidorah at 3:08 PM on September 27, 2020 [4 favorites]


I think they made a mistake by calling Russian bots "trolls", but I do believe there is an important difference between "Americans I disagree with" and "automated disinformation agents of a hostile power".

Also I was able to get 7/8 just by usernames. Bots have bot names. Firstname-stringofnumbers and Firstname-abbreviatedlastname are bot names. Don't tell the bots we know this.
posted by agentofselection at 5:50 PM on September 27, 2020 [5 favorites]


I got 7/8 and tripped up on chemtrail dude. In addition to the stuff they mention in their analysis I tend to look at join date and follower count for a quick sense of who is genuine and who might not be. Someone who joined last year and has tens of thousands of followers is an automatic red flag for me - they might be really good at brand promotion or have a publicist but probably not.

Chemtrail guy was a hard one that I vacillated on for a few minutes, mainly because I didn't know what to make of the gaps in his posts - that was down to my poor understanding of how Facebook works. I thought he was just posting to a private group and that's why I couldn't see those posts.

There's a website where you can plug in a twitter handle and get an estimate on how likely it is to belong to a bot or troll. I wish tools like that were built into social media interfaces, but it would probably just end up being selection pressure to make fake accounts more convincing.
posted by um at 6:49 PM on September 27, 2020 [6 favorites]


Got 7 out of 8. Still think chemtrails guy was a troll - okay, maybe not a bot account but troll nevertheless.
posted by M. at 8:51 PM on September 27, 2020


the fake accounts shown here tend to have fake-account-y names and profiles but honestly content-wise plenty of these are indistinguishable from dozens of accounts that probably do have legit users behind them

I mean, they are pretty much just political meme repost machines.
posted by atoxyl at 10:09 PM on September 27, 2020 [3 favorites]


I guess what I'm saying is - in context of the exercise they are presenting it's not hard to spot the fake accounts based on the criteria they are trying to teach (e.g. accounts that do not evince a real personality or inner life). But I question whether applying those criteria to a random sample of social media would not produce a significant number of false positives.
posted by atoxyl at 10:13 PM on September 27, 2020 [1 favorite]


As the analysis points out, when social media activity is motivated by real passion that usually results in the person behind the account exposing themselves a little bit - people who are genuinely committed to a set of values can rarely help themselves from putting more of their personality into their activism than they had planned. Trolls and bots, being by their nature insincere and acting in bad faith, never slip up like this because there’s nothing to expose.
posted by um at 11:34 PM on September 27, 2020 [3 favorites]


I think they made a mistake by calling Russian bots "trolls", but I do believe there is an important difference between "Americans I disagree with" and "automated disinformation agents of a hostile power".


Not for a lot of us sitting in other countries! I guess the automation is the difference? But a lot of these Russian troll farms aren't really that automated - there's still a person writing most of the posts, they're just being pushed to dozens of accounts each time.
posted by Dysk at 12:34 AM on September 28, 2020


Also, count me in to the group that believes that effectively all these accounts are trolls. I don't really care what is in somebody's heart of hearts, and it is literally impossible to know anyway. That leaves us with their words and actions to judge people on, and if it talks and posts like a troll, it's a troll.
posted by Dysk at 12:45 AM on September 28, 2020 [2 favorites]


you really think the professor from Rutgers is a troll?
posted by um at 6:03 AM on September 28, 2020 [4 favorites]


By the end I wasn’t even reading the tweets, I just glanced at the top of the profile and could tell whether or not it fit their criteria for “troll” (using scare quotes because I wish they had chosen a word that didn’t already have a different meaning). I can’t quite put my finger on why that bothers me — maybe it indicates their criteria is exceedingly narrow? Anyone who doesn’t fill out their profile or who doesn’t post about local events is a disinformation agent now?

And as others pointed out, is there a meaningful difference between a paid provocateur who tweets 100% terrible stuff and a person who tweets 90% terrible stuff? “Oh this person wants me and my family to die, but they’re not being paid to say that — they really believe it. Guess I better engage with them.” No thank you.
posted by Tehhund at 8:48 AM on September 28, 2020 [2 favorites]


The thing is, a lot of the risk doesn't come from people who post (what you think of as) terrible stuff. Some of the risk comes from Russian bot accounts who are posing as people with whom you basically agree, and who are trying to push you to accept more and more radical positions. So, for instance, they will pose as people on the left and try to convince left-leaning Americans that they shouldn't vote for Democrats, because the lesser of two evils is still evil. There are definitely some real people who think that, but Russian bots are working hard to amplify and mainstream that point of view. I actually think that the ones you need to look out for are the ones who are posing as your friends, not your enemies.
posted by ArbitraryAndCapricious at 8:59 AM on September 28, 2020 [6 favorites]


As the analysis points out, when social media activity is motivated by real passion that usually results in the person behind the account exposing themselves a little bit

I mean, sometimes, but some of these are basically low-tier Instagram meme accounts.

(Of course, the genuine ones might expose themselves when they decide to take a shot at making money.)
posted by atoxyl at 9:48 AM on September 28, 2020


...is there a meaningful difference between a paid provocateur who tweets 100% terrible stuff and a person who tweets 90% terrible stuff?

If you're trying to gauge the zeitgeist or get a sense of how divided and tumultuous the culture actually is, I'd say, "yes." If you're trying to decide whether or not to argue with someone online, then, no, certainly not.

Even just getting some explicit practice at recognizing divisive, provacative, unproductive emotional tactics -- whatever their immediate source -- and discounting them for what they are, is probably good for our brains. I thought this was a good idea, well implemented, but like everyone I wonder if it will reach the people who need it most.

I did very poorly on the quiz, but then, I'm actually collective of Estonian teenagers #naturalpenispillz #organic4life #chemtrails
posted by Western Infidels at 10:09 AM on September 28, 2020 [3 favorites]


Very good. Made me second-guess some accounts I follow. I appreciate that they pointed out how these trolls serve to dehumanize and divide people along multiple axes.

I think a lot of progressives believe that these trolls are only targeting conservatives. I run in progressive circles and many of my friends regularly share fake news and misinformation. It's really disheartening and I don't know how to address it without coming across like I'm discrediting or invalidating my friend at the same time...

NYT The Daily podcast recently aired their 2015 piece on The Internet Research Agency. I tended to think of bots as just obvious re-tweeters but the extent of what's being faked and the intent behind it is really alarming and eye-opening. It goes into the fake Louisiana plant explosion alluded to in this quiz. It's especially chilling to open the article's comments and see how many evident trolls popped in. It made me doubt much of what I read online, which is part of their intent. Article / podcast.
posted by Emily's Fist at 12:41 PM on September 28, 2020 [3 favorites]


The Mike Adams one threw me for identifying "sources of harmful misinformation" and not specifically "trolls," yeah.
posted by whuppy at 2:50 PM on September 28, 2020



One of the takeaways from this is that one strong characteristic of 'troll' accounts is that they list no personally identifiable information - like their employer, job position, place of residence, or first and last name (nor do they even unintentionally leak that information through their tweets).
I've done the same on Twitter as well but I'm uncomfortable with it because it means always give a second guess a user's authenticity and it decreases trust of people who for a variety of legitimate reasons (personal preference, security concerns, people who may not be public with their sexual or gender identity ) do not want to share that information.

(6/8).
posted by fizzix at 4:05 PM on September 28, 2020 [1 favorite]


I think a lot of progressives believe that these trolls are only targeting conservatives. I run in progressive circles and many of my friends regularly share fake news and misinformation. It's really disheartening and I don't know how to address it without coming across like I'm discrediting or invalidating my friend at the same time...

Yes, this is why I think identifying bots and trolls as early as possible matters. People radically underestimate their own susceptibility to the pressure that these activities exert, and refuse to admit they could be influenced. But we were, and are.

... I'm uncomfortable with it because it means always give a second guess a user's authenticity and it decreases trust of people who for a variety of legitimate reasons (personal preference, security concerns, people who may not be public with their sexual or gender identity ) do not want to share that information.

The thing is, I'm not sure that this hypothetical person who simultaneously engages in tightly-focused messaging while at the same time maintaining unshakeable operational security discipline with regard to their personal life can actually exist.

Eventually even the most passionate single-issue activist is going to slip up and re-tweet a picture of a cat, or post their own picture of a landscape they visited recently, or write a paragraph bemoaning the closure of their local cafe, or mention apropos of nothing that they are having a good or bad week. Over time you would get some sense of the human being behind the account. I think the only accounts that don't do this are either bots (which are constrained by their programming), or paid trolls (who are required to inhabit very specific roles).
posted by um at 4:29 PM on September 28, 2020 [2 favorites]


you really think the professor from Rutgers is a troll?

He's going on twitter to post things to garner a strong emotional reaction, written in such a way as to maximise the strength of reaction, from both people that agree with him and don't.

Like, I agree with basically everything he posts, but that way of going about engaging with the Internet? That's what trolling is - looking to get a rise from people.
posted by Dysk at 12:25 AM on September 29, 2020


Like, the feed is full of gotchas. Good gotchas, but still - there's a reason we generally consider just throwing things out to make others look bad isn't considered good faith engagement here on mefi, for example. And I get it, racists don't get out deserve good faith engagement. Totally fine. But that still means you're doing some bad faith engagement, purposefully riling them up. That is trolling.

It's righteous trolling, but still. It's fundamentally the same tactic - win the argument by getting your opponent apoplectic.
posted by Dysk at 12:28 AM on September 29, 2020 [1 favorite]


I got 5/8, which was 1 or 2 lower than I thought I'd get. I've been online a long time! One of the first things I ever did online was get banned from a CompuServe "CB" chat channel for griefing in some way! 1982ish! I've polished both sides of the fuckhead coin and I've prided myself for a pretty good ability to call out bad faith actors. It's a little humbling to have it measured!
posted by rhizome at 11:40 PM on September 29, 2020


« Older 🐭🥇   |   Inside Ebay's Cockroach Cult Newer »


This thread has been archived and is closed to new comments