I tend to go with the rule of thumb “could a bunch of humans come to a consensus on this judgment?” If not, it’s not for bots in the first place. AIs are still useful in those edge cases to assist a human editor by saying this language might be divisive for example prompting them to change it. But as for uncovering intent in human communication? It’s the wrong tool being shoehorned into the wrong problem.
17 Likes
Absolutely! Plus people acting in bad faith to the point of setting up accounts years before they plan to use them so they can surface with the appearance of being a long time member are not stopped by a warning that what they are about to post might be against the guidelines; they know it!
In fact, thinking about it, it could be used as a tool to provide instant feedback - helping them to craft a more duplicitous post! 
11 Likes
The benefit of the AI might be to provide an initial assumption of bad faith where a human moderator feels he cannot. Certain phrases or structures of discourse can provide clues to say “take a closer look at this one”.
The problem, as you note, is that the most toxic and persistent tr0lls here have always been the ones who know how to game the system and the community culture. They do so not only through establishing submarine and alternative accounts and long tenures but also through comments carefully worded and constructed to skirt the line of the rules. Only humans can detect that last tactic at the moment, which is why the Leader trust level is a more powerful aid to the moderator than AI will be as the technology currently stands.
It’s all too possible. I don’t think we need to go into the details of how. One thing mitigating it is the twisted pride a lot of tr0lls take in hand-crafting the turds they drop. If they get caught using mechanical aid they’ll be exposed as being more pathetic than they already are.
7 Likes
That was the first thought I had, as well. One could easily end up training the bad actors to be harder to detect.
6 Likes
Like…avoiding the word filter? 
8 Likes
I’ll readily admit that’s amateur-hour stuff. I’m mainly talking about the ones who take advantage of the moderator’s remit to assume good faith by disguising their trash opinions in a certain kind of language. There were bad actors who persisted for years here on that basis before they were finally banned.
9 Likes
I know. I could not resist though - it was just sitting right there
10 Likes
I see an arms race coming.
The first thing that came to my mind when the current hype started was: oh no, someone will use this to do credible disinformation. The second thought was: oh shit, someone is going to use it for SEO. The third thought was: … it’s a fucking chatbot, everyone and their dog will flood the comments and piss on others for fun.
(Just for the record: I hate the term “AI” right now. It is so … regressive, I guess.)
8 Likes
A snark hunting AI? The Rev. Dodgson would be proud.
9 Likes
I’ve never understood that mindset. That is such a weird hobby.
5 Likes
Unfortunately, the Walrus and the Carpenter are going to own the AIs…
8 Likes
Cabbages and kings.
7 Likes
The AI needs to be designed to triple-check itself before coming to a conclusion.
"Just the place for a Snark! I have said it thrice:
What I tell you three times is true."
8 Likes
As always, there’s a relevant xkcd:
24 Likes
I’m already seeing so much b.s. in my newsfeeds, but at this time they’re fairly easy to spot because of headlines with poor grammar. It’s not going to be so easy to see in the future, I don’t think.
13 Likes
Hiya
I recently had a post flagged for commending the author for a great article and not simply reposting some twitter links.
I get Twitter is an information source, and BB has had long linked to its content.
What I don’t get is why BB continues to shine the light on the Musk/twitter dumpster fire while still directing traffic there?
If anything the Twitter links seem to have increased in the past few months - which seems odd. Anyone help clarify?
14 Likes
ok thanks.
In the grand scheme of things, having BB post links to Twitter isn’t going to make a huge difference if it lives or dies.
cheers
10 Likes
Oh, shoot, I’m going to have to ameliorate my language.
13 Likes
14 Likes
