Actually programmed accounts, which until very recently were limited to algorithmic responses like counting words, doing unit conversations, or giving random facts
Disingenuous actors seeking to influence conversations by using multiple fake accounts
The former is what some Lemmy instances “allow”. The latter is a complex moderation problem that isn’t really allowed so much as hard to differentiate from (dumb, wrong, whatever) organic responses. And with LLMs now may be able to be done algorithmically as well.
Not that easy to block them. Many of them are part of sophisticated overseas influence operations that may be using local proxies. A couple years ago it was easy to spot them, now it takes work. When you’re dealing with chat traffic on this scale…it’s a time-consuming job.
It’s up to each instance admin team whether to allow bots (where they can identify them) and whether to defederate from instances that do allow them (again, where they can identify them)
I was surprised that Lemmy allows bots. Terrible idea.
There are two types of things called bots:
The former is what some Lemmy instances “allow”. The latter is a complex moderation problem that isn’t really allowed so much as hard to differentiate from (dumb, wrong, whatever) organic responses. And with LLMs now may be able to be done algorithmically as well.
Not that easy to block them. Many of them are part of sophisticated overseas influence operations that may be using local proxies. A couple years ago it was easy to spot them, now it takes work. When you’re dealing with chat traffic on this scale…it’s a time-consuming job.
Meanwhile, your account is less than one day old.
I have seen no evidence to support these claims.
It’s up to each instance admin team whether to allow bots (where they can identify them) and whether to defederate from instances that do allow them (again, where they can identify them)
It’s up to each instance