I’ve set registration to Require application again, don’t like this bots creating accounts.
Also I’m running into email limitations. Using the Sendgrid Free plan of 100 emails per day, but if a couple of member have email notification on, I’m reaching that 100 very quickly. How do others handle the high email traffic ?
It serves the same point it served on Reddit. There are huge bot farms dedicated to shaping the social media landscape. There is a new social media “boom” so the bots want to get in on the ground floor.
with what purpose ? posting stuff, or some other use ?
It’s a variety of purposes, but in this type of social media space it’s usually to push specific agendas, often political. The accounts will often engage primarily by “reposting” comments to give a profile appearance that looks like a person. Then they engage in manipulations of conversation in specific ways. For instance, they might play “devil’s advocate” in left-leaning political spaces to plant certain view points. One of the quirks of the human brain is that we tend to believe things we read over and over, regardless of the source, because volume breeds validity. It’s a part of being a social species and being hardwired to be part of the herd.
If you hang out on social media and you read over and over, from multiple posts, that someone is a war hawk, that becomes part of your “sense” of them. Often these conversations are subtly shifted just by posting opinions with little to no substantiation. Your brain starts to think “wow, there are a lot of people saying that guy is a war hawk” and, since we’re wired to conform to the group, you wrap that into your total opinion of the person. You’ll also see just general “shaping” posts without a clear angle where the intent is to shape your perception of a public figure negatively or positively.
Think about your opinion of a public figure. Just your sense of who they are. Do you actually know facts that guide what you feel, or do you just have a sense of them based on all the things you “know” about them?
All true. Don’t forget that they can be used to vote, meaning content supported by the bot operators rises and content they don’t want you to see is effectively hidden.
Advertizing, astroturfing, shaping perception via upvote/downvote. Remember that time reddit posted user statistics like pornhub does, and it turned out the most “reddit-addicted” city in the world was Eglin airforce base, Florida?
Concur. Instance admins need to be stamping these out with prejudice.
Future spam accounts in the making. They are building large bot armies to either flood instances with their crap, or use them for vote manipulation, or a combination thereof.
I read somewhere that you can add a captcha to registration, maybe that’ll help filter some of the bots out?
Try enabling the captcha in your instance settings, it might help a bit as well!
While looking for a way to enable captchas I found out that they will remove that feature from lemmy: https://github.com/LemmyNet/lemmy/issues/2922
That’s unfortunate. A lot of bots have beaten them but it still stops a lot of low quality bots
Wouldn’t it help to force registrations solve an expensive proof of work on the client side?
That would probably only slow them down. And since their main goal is to create a ton of working accounts, speed is probably not that important to them. On my own instance the bot-registrations stopped immediately after I enabled the required application.
Captchas are mostly useless nowadays against any well designed bot. The only ones that sort of work are those ridiculously hard ones like several page math problems. And at that point a large portion of people can’t solve it either.
They also tend to be full of trackers and are terrible for privacy. Plus they’re used to train corporate AIs.
It can help temporarily, because a lot of bots don’t bother with captcha solvers, but all the spammers need to do is include a pre-made solver library and the bots are back.
Good idea thanks :)
Probably potential spam accounts.
My modlog got spammed today by some script banning a lot of people for spam automatically. It’s getting wild. This was inevitable with the growth of Lemmy as a platform though.
Mine too I see now. You can only ban a user when they have posted something right ? There is no list for admins of all users registered ? Maybe I’ll do it direct in the database…
Just did that for my instance. Luckily it’s fairly easy. Since there were no human registrations after those bots I just looked up the lowest bot id in the
local_usertable which was9and then useddelete from person where id in (select person_id from local_user where id >= 9);. Thanks to the references and triggers on the table, the rest of the database was also cleaned up by that, including thelocal_usertable.For me it were the last 30 or something entries in local_user. They all didn’t have email_verified. I access the db through Postico. Also removed all other older accounts that didn’t verify their email address
