Excellent. This was a community suggested option to help. Especially running up to the election when shenanigans are accelerating.
25 Likes
Having an AI to assist the moderators may not be a bad thing if it were trained well on the corpus of BB posts (both the good, the bad, and the ugly that were flagged and those flags upheld.) It could flag things that may have flown under the moderatorsâ radar for them to take a closer look at. But Iâd be cautious about letting the AI take action on its own. Maybe in the most egregious cases (some sort of denial of service-like attack) Iâd give it a bit more slack in its leash, but generally speaking Iâd want to have human-in-the-loop moderation day-to-day. [And I donât think, based on this discussion, that anyone is proposing fully automated moderation at least for right now.]
5 Likes
have to sit and think over an âA.I. trained on the corpus of BB postsâ (by which iâll read âcommentsâ). that would be one curiously stressed out âA.I.â indeed [wink] (âsâgotta be better than most 'A.I.'s out there which are mostly trained on reddit commentsâ)
6 Likes
Iâm fairly sure Iâm not AI, so two out of three isnât bad.
19 Likes
Of course, but can you really ever know?
20 Likes
Perhaps not one AI, but maybe a few. One per category? Or one per âsufficiently largeâ category? In decreasing order of # of topics boing (110k), general (4.6k), and wrath (2.2k) seem big enough to support an AI each while meta (834), links (652), games (576), craft (170), and Offworld (20) may be small enough and/or targeted enough to not need additional monitoring. [It looks like Offworld is dormant or defunct?]
boingâs corpus, sorted by # of replies in each topic, does seem like it would be a little bit spicy. The top 3 deal with Russiaâs invasion of Ukraine, January 6, and Trumpâs election. But thereâs probably a lot of posts that were in those topics that no longer show up that would give the AI some good guidance on what to flag.
wrathâs corpus is heavily politics focused based on the list of topics sorted by # of replies, which would likely skew its recommendations. But again, like boing it probably would have a lot of stuff to learn from.
general would be a mixed bag. A bunch of coronavirus topics, âOdd Stuffâ topics, âHappy Mutants food and drinkâ topics, and âGood (Encouraging) Stuffâ topics.
4 Likes
We need to add that to the list of things that start arguments on bb.
15 Likes
Paging @NukeML!
15 Likes
Good luck with asking the mods about that, theyâve had less than a day in the job and youâre planning their replacement.
16 Likes
Maybe theyâre too old? And should be replaced with a giant robot baby?
What could go wrong?
13 Likes
18 Likes
Fine. A giant, drunk robot baby?
19 Likes
Good. So that leaves all others in the community to wonder about, such as @Brainspore who I believe is an actual AI final âgoldâ version. All fingers and toes, present.
7 Likes
âMan. Woman. Person. Camera. Tv. Giant. Robot. Baby.â
âcognitive test results, probably
source
[probably. ish.]
13 Likes
So weâre not going to be told the new modsâ original identities? Booooo!
4 Likes
Replacements, no. As I said, I wouldnât trust AI unsupervised to moderate BBS. Itâs difficult enough for human beings to interpret what other human beings write sometimes (what with the written word not providing some of the context clues that help out when dealing with someone face-to-face like body language, the tone of a personâs voice, etc.)
Thinking about possible tools that might be able to help them with their new responsibilities by providing a first pass filter or a last filter to catch issues that may slip through the cracks, yes.
3 Likes
Nice Kitty! Good Kitty!
Do please give Ptah, Nefertum, and Maahes my love.
15 Likes
Of course!
13 Likes
I didnât realize anyone else even worshiped @Community_Mod_1!
21 Likes
I love and appreciate all our mods, mortal, divine, and divine-adjacent!
12 Likes
