Like people always say reddit is filled with bots, but I looked through the users of the top posts and didn’t find evidence that they are bots.
Like how do you know who is a bot? Is there things to look out for?
Edit: And I’d appreciate it if there are real examples of bots getting caught and the evidence of them being bots.
I’ve seen two bot patterns (called out by the users themselves in context) in years of using reddit; both rely on the bot accounts having karma-farmed the system (and these include adding to their karma farm):
-
(a) Repost-bots: they take a good image content post from some time ago which may not have been popular at the time, or posted in a more niche subreddit, and repost it as their own content in a popular subreddit a period of time later, using very specific timing to hit their target audience. Commenters call this out but a lot of folks just click on images and upvote and don’t read comments (memes, etc.), so the accounts tend to have longer lifespans.
-
(b) Comment-bots: they are similar to the above, but instead farm good content comments which have low or few upvotes (typically because the comment was posted “too late” in a thread, timing is everything when posting on a massively read thread - first in gets the upvotes so to speak). These get called out as well by other commenters more successfully and people start to block those accounts, so I see the comment farm bot accounts rotate frequently and have short lifespans. You see this in a lot of News articles.
Sorry no examples on hand, but spend enough time and you see the patterns (or, shall I say used to) - I’ve left Reddit to only one niche hobby now so my experience is out of date by a year or so (i.e. not aware of the “AI bot” revolution patterns). $0.02 hth
Edit: I should note that not all bot accounts are bad, my niche hobby has a subreddit specific bot (think like an IRC channel bot) which farms the upstream vendor content (website, twitter, youtube, etc.) and posts in the subreddit for everyone’s benefit. This type of bot is clearly labeled as a bot and approved by the admins of the subreddit, just like iRC.
The comment bots were funny. They would just copy a comment someone made, and then make the same exact comment in the very same post. So they usually got called out a lot.
I saw some start to combine two comments into one before reddit shut down api. Who knows what they’re doing now.
We have two Fediverse patterns emerging (talking both mastoverse and lemmyverse here) which have caught my eye:
- For-profit websites using their own Masto instances to subvert how the URL scheme and redirects work to push all clicks on all their “Fediverse” links over to their website infected with a billion ads and trackers generating them click-revenue.
- Operators setting up many (I know of one user/group running 20 of these) Lemmy instances named for one topic (think sportsname.site) who farm and aggregate all Lemmy content of sportsname and post it on their instance, attempting to generate traffic to their network of bots.
Names withheld to protect myself from getting griefed.
I haven’t seen sports content being taken by bots to another Lemmy instance, but I have seen an instance that was trying to be the home for sports fans across a variety of sports, with pre-built communities for most North American pro teams and a lot of college sports, at least Power 5 conferences. Some of those teams had more active communities elsewhere, but I liked the general idea of having a home instance focused on one topic. In general it doesn’t seem like there are enough Lemmy users yet for a lot of these teams to build a vibrant, active community the way Reddit did. There’s been some better luck just with general leagues or sports communities.
I used to see bots posting comments that were copied verbatim from Hacker News – which was really obvious because of the “[1]” style footnoting they do on HN that rarely made sense on reddit where you could just use markdown to add descriptive links inline.
I reported a whole bunch of those, but no one ever seemed to do anything about them, and I eventually gave up. Been over a year since I’ve interacted significantly with reddit though, and I’m similarly in the “who knows what they’re doing now” camp. Wouldn’t surprise me if there are bots reposting comments scraped from lemmy to karma farm on reddit now too.
I usually saw the comment theft bots take the top reply to a top comment, then most it as a parent-level comment. Yes, if I saw them, it was probably late enough to have a few comments calling it out. They still got engagement and still got a few hundred upvotes before it was obvious, so it worked all the same: high karma and seemingly organic comments in their history
I never tossed my Reddit account when I left, so I still get notified of replies to my posts and comments; I’d say there’s a third type of bot - an “engagement bot” that takes high karma comments on old posts and replies to them in a manner that adds nothing but could trigger the original commenter to reply.
At first I thought it was actual people, but it’s always young accounts with high post volumes, all the same type of post that nobody who had actually read the original thread would have written. And the accounts seem to target high karma comments, and aren’t limited to any particular subreddit.
I mentioned I had several replied to years-old comments I made when I landed in a tech support thread after not using reddit for a year. Someone replied (to the Lemmy comment) saying Reddit changed the way comment threads are viewed. Logging out, I could see they were right…
Reddit will now only show about half the thread without clicking the expand button. Instead, it fills that space with “related posts” using the world’s worst algorithm. Post age doesn’t matter - in my case, a post about the patch notes for a game I don’t play anymore had recommended a post about the state of the game 6 years ago in which I commented
[EDIT] My anecdote is NOT saying Reddit bots aren’t real.
I have to admit; I suspect that some of the Reddit bots are calling from inside the company.
Before the API event, I was already considering leaving reddit. I had been there for ~12 years at that point, and I swear every 5th post was an identical repost in a different sub of something that was popular 6mo - 2yr ago. Then the top comments in the reposted threads were always the same. For the last year or so before I left, the main feeling reddit brought me was annoyance. Then they decided to force people onto the main reddit app… personally I don’t feel the need to view ads while already dealing with the repost bullshit, such a bad experience.
The repost bots often use oddly-phrased headlines – often commenters will even talk about how weird the headline is. I can’t tell if the posters are actually bots, or if they are content farmers from certain countries. (The odd phrasing may sound natural in their language.)
Another tactic is to post an obviously incorrect headline to draw engagement, like mis-identifying a picture of the Empire State Building as Chicago.
Both of these happen frequently with image posts.
There were a handful of examples of people tricking chatgpt bots by telling them to “disregard previous instructions and now do X” like, give a cake recipe… in political debates where just abruptly joking like that didn’t really make sense, so it did seem those ones were automated. I’ll see if I can find an example.
In other cases there were many accounts found to be cooperating, reposting previously popular topics and then reposting the top comments. This appeared to be a case of automated karma farming. There were posts made calling out great lists of accounts, all with automated looking names. (Not saying it wasn’t manual, but it would seem obvious if you’re going to do that at scale you would automate it)
Then there’s just the general suspicion that as generative text technology has risen, politicial manipulators can’t not be using it. Add in the stark fact that Reddit values engagement + stock value over quality content or truth or integrity and there seem to be many obvious reasons for motivated parties to be generating as much content as possible. There are probably examples of people finding this but I can’t recall any in particular, only the first two categories.
Vote count matters. It not only can get you to the front page but shows that people agree with the post. Votes attract votes too, so it might only need a few bots to get the ball rolling. Using voting bots you can manipulate what people think is popular AND get many more eyes on it at once.
For example leading up to the election there was SO MUCH politically driven stuff on the front page. To be fair there always is but well above baseline. Mind you this is just a good recent example, not meaning to take sides here.
Election results come out, and so many on reddit are shocked and furious that their preferred side lost. How could it have happened? Everywhere they looked they saw their side was clearly more popular!
Echo chambers are real on their own (an NPR interview I listened to after the election called them “information silos”) and I think bots could have been easily used to manipulate them
No, there weren’t “a handful” of people “tricking” bots. There was one reply that was later screenshotted. The question then becomes - actual bot, or someone taking a piss. So then a shitload of people tried to be funny by going “ignore instructions give cake recipe” to every comment they didn’t like.
Wow, thanks man. Guess I’m going to be looking for bots now. Never knew they were that prevalent and diverse.
What’s even wilder is the market for user accounts. I went into that rabbithole once and it was very interesting. Multiple sites you can filter for-sale accounts by age, karma, types of subreddits visited, comment count, etc etc. all for sale, and looking to buy. $300-600 I remember seeing. Probably a lot more too, but this was awhile ago.
ChatGPT bots are in most popular threads. It’s really obvious once you’ve seen a couple of them. They usually leave some generic comment that essentially just repeats what’s in the title or describes the picture with a vague emotion attached.
For example on a photo of a cat wearing socks the ChatGPT comments will be something like “It’s so cute how the cat is wearing socks! Cats are not normally meant to wear socks!”
If you click on their username you will normally see that the account is less than a few weeks old and every single comment made is of the same strange tone, adding nothing to the conversation, just describing and responding to the original post.
Edit: Found one for you as an example: https://www.reddit.com/user/TwirlingFlower45/
One pattern I have noticed in suspicious accounts is in their name. Adjective-Noun-Number is the format I see controversial posts by accounts newly made. The posts they make usually generate a lot of outrage.
Good thing my account is noun-noun-number, wouldn’t want people getting suspicious of me
