hey everyone. if you want to post links or discuss the Reddit blackout today, please localize it to this thread in order to keep things tidy! Thanks!
hey everyone. if you want to post links or discuss the Reddit blackout today, please localize it to this thread in order to keep things tidy! Thanks!
I was wondering about how much of reddit activity is bots. Are bot posts/comments possible with lemmy?
It’s impossible to tell how much of the posts are bots but if you look at the default subreddits I’d say that it’s definitely noticable.
Reddit is a popular grey marketing area - be it shilling products or political propaganda. Lemmy/Federation hasn’t reached enough mass for this to be a real problem yet but it’ll happen eventually and Lemmy is an easy target right now. I used to work in bot detection area and modern, well made bots (the ones you should worry about) are essentially indisguishable from real users but script kiddies can be an issue too.
The only real way to fight bots is to reduce the incentive which is more of a cultural thing - people have to call out shills and a more transparent platform definitely helps.
Sorry for the rant but I think bots/spam will play a big role if federation ever reaches the point where there’s enough eye-balls for shilling to be valuable.
Good perspectives, I totally agree it’ll be a big challenge here as the fediverse gets bigger. Mastodon’s already had a couple of waves of spam - Brian Krebs had a good article a few weeks ago
I dunno, there’s a particular instance here in the fediverse that seems like an alt-right bot operation. I don’t want to mention it directly to avoid the ire of said community, but once you find it your Head might explode.
I’m certain lemmygrad is a propaganda operation by China or Russia. Well at least I want to believe that people aren’t that toxic and ignorant willingly.
There has to be a way to limit these accounts. Perhaps ban accounts with inhumane levels of activity? If possible, this would be awesome. I think that’s what makes a platform for me - interacting with real humans on topics of mutual interest. It would be such a bonus to have a feed not dominated by bot reposts and inevitable drivel.
Behavior analysis is definitely a popular way to handle this but it’s very difficult and resource intensive. Major issue is false positives as users are hard to predict in flexible environments like forums - is this person interested in all Honda posts as they are researching their new car purchase or they’re shilling for Honda?
It’s tough.
Yes, there’s nothing stopping someone from creating bot accounts on Lemmy and pretty much every other Fediverse platform. Even if APIs are restricted, they can just parse HTML instead (though that’s a bigger pain in the ass). This is an area where the decentralized nature of the Fediverse works in our favor, though, as it inherently limits reach and discoverability (thus minimizing the benefits of doing this). For example, Mastodon’s flagship instance (mastodon.social) had a spambot problem not too long ago, so what happened is when other instances noticed this spam wave, they limited/defederated with mastodon.social and the problem was solved on their end. The host instance can temporarily close off sign-ups to prevent new accounts from being made. Every other instance can control federation to effectively quarantine the spam problem.
Interesting. My initial thought was that botting on the fediverse might be easier than on a centralized site, since you can target your bots to several instances at a time. If one of them catches on, your bots can still function because they aren’t being banned across all instances.
What you say about being able to quarantine and cut off instances that are being spammed makes a lot of sense though. I’m glad that at the least there’ll be some new challenges for bot authors.
Ultimately though, the relatively small size of the communities is our best defense against bots at this point. Lemmy/kbin just aren’t great targets yet because bots won’t reach enough people to achieve their creator’s goals.
Probably, but it wouldn’t be as easy since each instance has different design, admission criteria, and access to other instances. A bot would have to be based on a given instance, so a botnest instance could be cordoned off. If it becomes a bigger issue, each instance may have their own way of dealing with it so the bot would have to be capable of not only negotiating the above obstacles but also how to get around a variety of different ant-bot measures.