Moderating is so hard it will probably the first application for AI with generalized intelligence.
Personally I think it feels kinda dystopian for an AI to be deciding what human beings are allowed to say; _especially_ if it's highly effective. There was a comment from another user I favorited a while back which reflects my thoughts on the topic of AI-driven moderation pretty well[1]: > I find censorship via algorithms extremely scary. There will be false positives and false negatives and there is absolutely zero…
The issue with that is we have already been using algorithimic filtering for a very long time in internet terms. Spam filters are ubiquitous and help keep out the infamous online Viagra sellers and comment posters in check. It is a very limited subset of course.
Humans already have generalized intelligence and this is the result of that. All generalized AI will do is help us systematically make the same mistakes at scale.
It sounds weird but don't underestimate the positive value of making the same mistake at scale instead of a bunch of different ones. That is essentially the model of laws and bueracracies and as much as the latter draws complaints it works far better than literal fiefdoms. Even when the holders were trained essentially from birth.
That’s an interesting point. However, AIs are much more opaque in their logic than laws are. I’m not so optimistic that the public visibility will be quite the same.
"The police" would consist of government bureaucrats and cancel mobs. Facebook can be and is being reported to them with the full expectation that they will ultimately be able to force Facebook to censor in different ways. > One Mafia family has functionally unlimited power, and verifiably censors the two other groups worldwide. There's a media blitz about Facebook precisely because Facebook doesn't have unlimited po…
>>One Mafia family has functionally unlimited power, and verifiably censors the two other groups worldwide. >There's a media blitz about Facebook precisely because Facebook doesn't have unlimited power, B-b-but I thought Facebook wasn't one of the Mafia families? I think you're just saying words for funsies. >The media, the government, and the mob can fight Facebook on terms that give them a meaningful chance of winn…
>B-b-but I thought Facebook wasn't one of the Mafia families? I think you're just saying words for funsies.
The analogy is that one Mafia family can turn in another Mafia family. Facebook has always been a Mafia family in this analogy.
>>One Mafia family has functionally unlimited power, and verifiably censors the two other groups worldwide. >There's a media blitz about Facebook precisely because Facebook doesn't have unlimited power, B-b-but I thought Facebook wasn't one of the Mafia families? I think you're just saying words for funsies. >The media, the government, and the mob can fight Facebook on terms that give them a meaningful chance of winn…
>B-b-but I thought Facebook wasn't one of the Mafia families? I think you're just saying words for funsies. The analogy is that one Mafia family can turn in another Mafia family. Facebook has always been a Mafia family in this analogy.
Whatever you say. Facebook, the right, black people, and the LGBTQ+ are all on a level playing field, and everybody's bullying Facebook. It's a terrible, incoherent analogy, entirely unable to comport with the actual relative power of these groups. Which is why you're here shuffling goalposts into entirely new sports fields with me instead of engaging with others elsewhere who brought up inconvenient factual information about what's actually going on; you're unwilling or unable to answer any questions about the New Zealand shooter, for instance, because that sort of real-world example exposes the inanity and indefensibility of your position. You're unable to acknowledge that one of the groups provably, successfully censures the others, continuously, worldwide.
Personally I think it feels kinda dystopian for an AI to be deciding what human beings are allowed to say; _especially_ if it's highly effective. There was a comment from another user I favorited a while back which reflects my thoughts on the topic of AI-driven moderation pretty well[1]: > I find censorship via algorithms extremely scary. There will be false positives and false negatives and there is absolutely zero…
The issue with that is we have already been using algorithimic filtering for a very long time in internet terms. Spam filters are ubiquitous and help keep out the infamous online Viagra sellers and comment posters in check. It is a very limited subset of course.
I have no issue with using algorithms to filter machine-generated content, only human-generated content. I realize from a practical perspective that can be a difficult distinction to make, especially for an algorithm. In principle though I think it's an important one.
A charitable interpretation would be to understand the "arabic" as in "The arabic script", that is: the alphabet. This is indeed the written representation of Persian and Urdu. It's a common confusion for some people to conflate a natural language with it's script. Compare with the similar confusion that conflates a programming language with either it's implementation or the snytax used to write it in text files.
By this logic Chinese, Japanese, Classical Vietnamese and Classical Korean are all sorta kinda just Chinese. Aside from being amazingly ignorant ( it can be argued Chinese dialects alone are more diverse than Romance languages) your not going to be able to have a Cantonese speaker moderate a Japanese forum. That said, why are we realistically expecting Facebook to be able to do all this. The moderation should really…
What the hell are you saying? Is it "the existence of state censorship is a good countermeasure to online speech when applied at the platform level"? Because that would be sneaking some pretty authoritarian shit into your opinions. I want facebook to go away too but unilateral action on the part of states isn't going to do it (no matter how influential you might want China to be). Now, some kind of international coalition to support fact based information via sanctions (otherwise you trade one kind of propaganda to another) might help.. guessing big red not into that idea tho.
Can you imagine what a dystopian sci-fi hell hole the metaverse could be with facebook at the helm? They can't even moderate a website.. the hubris is pretty amazing.
i mean i get the hate, but categorize moderating facebook as 'can't even moderate a website' is pretty outrageous
I meant to stress that their metaverse vision seems to be a much more massive endeavor and that it would probably be a good idea to learn to walk, or moderate their website (or whatever you want to call their tech that interfaces with their users through a website and app) before they try to run and create something that they most likely will also not be able to control effectively.