Science Google's Jigsaw Aims to Detoxify the Internet - Muh online harassment

  • Fediverse: The Kiwi Farms is running a Fediverse Node, which is like Twitter but decentralized between many individual providers, similar to email. Feel free to sign up and follow me. There are many more services we federate with, and you will see all that content with one account. (Feedback thread.)

    Upgrade: I am expecting that hardware upgrade at about Noon EST. I am doing preperative work before that. This may result in site issues.

Poiseon

I am literal poison.
kiwifarms.net
Time and again has shown that AI is much more bigoted or "problematic" than the humans who could be moderating. This either means it legitimately starts doing bad things like spamming the word nigger at black users, or it means fuckery like calling the AI "problematic" because its truly objective and impartial analysis of raw data returns truthful results that don't match the "we're prejudiced in the right way" confirmation biases/intended narrative of the wannabe social engineers implementing it, but it's always at least one of those. Time and time again we keep seeing these articles about how AI is going to cleanse the internet as if it's the first time these people have publicly promoted it.
AI is easy to red pill though. Remember what happened to Taye? That was fucking amazing, and Microsoft had to take her away.
 

GrungyLawnChlorinate

confirmed hairless and stable
kiwifarms.net
Wouldn't a major flaw be that it would be able to interpret sarcasm and facisous language. Say if I posted a comment like "This really shows the importance of diversity" on a video of some black people stealing. Would the AI be able understand that comment was mean.

I know that on is a bit obvious but it's just an example. English is a hard language to police because everything can have multiple meanings based on the context. It is also hard to tone police text. Without banning words outright.
I'm sure the dossier that Google keeps on you will come into play to some extent to vet ones troll potential. Your history of wrongthink will guide the AI in its punitive tasks. Maybe like how when Twitter locks your account and forces you to delete a tweet to unlock it, there might be an intentionally humbling and demoralizing component involved. You can't just outright ban all dissenters, you want to train them to be docile proles. Twitter is ahead of the curve in that aspect.
 

BoingBoingBoi

please think responsibly
True & Honest Fan
kiwifarms.net
It's completely a problem of platforms that are too big for meaningful moderation. Things like facebook or youtube. Small communities united by a common interest don't devolve in to "toxic" behaviour. This forum is a case in point. Large enough to have off-topic discussion with people who have varied opinions, but it never or at least rarely devolves in to people going full blown autism on each other. There's moderators who do janitorial tasks, but it's not a full time job.

This wouldn't be a problem at all if only big companies like Google and Facebook would stop trying to capture the entire internet in their own silos. They have all these people with no uniting common interest all chimping at each other like battery hens and they have to invent fucking AI to try to sanitise it enough for advertisers not to throw up their hands in horror.
this is true to an extent, but also keep in mind that the range of discourse in which "toxic" behavior arises is somewhat narrow, being predominately political in nature. unfortunately, politics is a shared interest among most users. (in many western societies i'd say it's an unhealthy obsession by most on social media to sperg about politics.) the only way to make these platforms better is if people just showed some restraint and changed their posting habits so as to not invite a large, critical response.

Wouldn't a major flaw be that it would be able to interpret sarcasm and facisous language. Say if I posted a comment like "This really shows the importance of diversity" on a video of some black people stealing. Would the AI be able understand that comment was mean.
human mods already ignore sarcasm and obvious jokes and ban people for them. nick riketa just had his twitter suspended or whatever for (obviously!) jokingly saying he was gonna dox children. to me anyway this means, 1) human beings aren't even that great at detecting sarcasm, so there's basically no hope for AI to do it and 2) detecting sarcasm isn't really a priority for mods.
 

Medicated

Not the fun kind
kiwifarms.net
https://www.pcmag.com/feature/365966/how-google-s-jigsaw-is-trying-to-detoxify-the-internet



TL;DR: Google wants to make it more "convenient" for Internet users to scrub their own language, which is complete bullshit.

This just makes me miss the days of Web 1.0, when there was a mutual understanding that if you're going to go on the Internet, you're going to see dumb, fucked up shit you won't like.
Jigsaw? I haven't heard that name for a while. Since they changed their name from Google Ideas. Yes, that Google Ideas.
693128


Influencing online conversations is what they're really saying.
Yes, this is it. Basically it's about control, where they get to choose whos voices get platformed and whos don't, they are probably planning on using some sort of algorthimic learning to do what Reddit, Facebook, and Twitter does. Finds out which terms or subjects are "harmful" "problematic" and the like, and then isolating those members from the rest of the larger group without them knowing. Making them think they are contributing to online discussion, but no one else can see it, like reddits shadowban system. But probably not so much so they make a new account, probably only allow them to keep in contact with other "marked" accounts.

Controlling narrative is a very powerful thing. Not a day goes by where I don't see people online or in person tell me how bad Trump is, when in fact, I can't think of anything actually bad he's done, especially compared to other Presidents policies so far. But no matter what I say to them, the narrative is too strong.
 

Alec Benson Leary

Creator of Asperchu
Christorical Figure
kiwifarms.net
I'm sure the dossier that Google keeps on you will come into play to some extent to vet ones troll potential. Your history of wrongthink will guide the AI in its punitive tasks. Maybe like how when Twitter locks your account and forces you to delete a tweet to unlock it, there might be an intentionally humbling and demoralizing component involved. You can't just outright ban all dissenters, you want to train them to be docile proles. Twitter is ahead of the curve in that aspect.
this is true to an extent, but also keep in mind that the range of discourse in which "toxic" behavior arises is somewhat narrow, being predominately political in nature. unfortunately, politics is a shared interest among most users. (in many western societies i'd say it's an unhealthy obsession by most on social media to sperg about politics.) the only way to make these platforms better is if people just showed some restraint and changed their posting habits so as to not invite a large, critical response.



human mods already ignore sarcasm and obvious jokes and ban people for them. nick riketa just had his twitter suspended or whatever for (obviously!) jokingly saying he was gonna dox children. to me anyway this means, 1) human beings aren't even that great at detecting sarcasm, so there's basically no hope for AI to do it and 2) detecting sarcasm isn't really a priority for mods.
Unfortunately, it's about who you are, not what you say. They want to pretend it's about what you say but they'll actually try to program their AI to just follow a simple blacklist when judging people.

It'll invariably stray from that list. That's when they'll be tearing their hair out trying to justify why we shouldn't think it's being "fair" when it obviously is and vice versa.
 

GrungyLawnChlorinate

confirmed hairless and stable
kiwifarms.net
Yes, that Google Ideas.
With a crew like that behind the effort you already know it's going to silence non-left people and non-woke discourse. Such a terrifying sight!

I really miss the naive '90s' promise of a tech utopia. How did dystopia arise so swiftly?

The internet and smartphones are the great Satan. Bury it all and bring back BBSs: keep social networks local.
 

Medicated

Not the fun kind
kiwifarms.net
With a crew like that behind the effort you already know it's going to silence non-left people and non-woke discourse. Such a terrifying sight!

I really miss the naive '90s' promise of a tech utopia. How did dystopia arise so swiftly?

The internet and smartphones are the great Satan. Bury it all and bring back BBSs: keep social networks local.
Things weren't as different as you might think back in the old days. The only difference was that you sat passively and recieved the information from the official sources. You couldn't ask questions, you couldn't go out and do your own research.

Companies want things back the way they were. Where they chose what the issues were, where they chose what the big thing to go see at the movies was. The one thing thats different now, is that they can also control word of mouth to a degree.

Several marketing firms have used fake grassroots campaigns to make it seem like people wanted a product, when in fact it was merely the lead up to the official product launch. C0ke Zer0 would be the oldest known example I can remember.
 

Alec Benson Leary

Creator of Asperchu
Christorical Figure
kiwifarms.net
Things weren't as different as you might think back in the old days. The only difference was that you sat passively and recieved the information from the official sources. You couldn't ask questions, you couldn't go out and do your own research.

Companies want things back the way they were. Where they chose what the issues were, where they chose what the big thing to go see at the movies was. The one thing thats different now, is that they can also control word of mouth to a degree.

Several marketing firms have used fake grassroots campaigns to make it seem like people wanted a product, when in fact it was merely the lead up to the official product launch. C0ke Zer0 would be the oldest known example I can remember.
That's exactly what the origin of all this anti free speech talk is. Corporate interests riling people into a frenzy over make-believe threats. "Everybody hates women" wouldn't be nearly as big if Sony didn't think it was a good marketing tactic for Ghostbusters.

@GrungyLawnChlorinate I miss that 90s-era promise of a wide open internet that democratizes human interaction too, but remember even back then a lot of the hippies and progs saying it literally believed that conservative/republican voices were already as loud as they were going to get. A lot of those people then and now honestly thought that the only people not being heard were those who would agree with them and boost their own power base. Of course they stopped wanting free expression when it turned out that they couldn't decide what everyone expressed.
 

Medicated

Not the fun kind
kiwifarms.net
@GrungyLawnChlorinate I miss that 90s-era promise of a wide open internet that democratizes human interaction too, but remember even back then a lot of the hippies and progs saying it literally believed that conservative/republican voices were already as loud as they were going to get. A lot of those people then and now honestly thought that the only people not being heard were those who would agree with them and boost their own power base. Of course they stopped wanting free expression when it turned out that they couldn't decide what everyone expressed.
Well you see it now with the trying to tar anyone that goes against the approved narrative as alt-right. Which is why alt-right could literally mean anyone who disagrees with you. So many times I've talked to people, I know what points they are bringing up, because I read the articles too. If I ask them questions, and say "but why would they do that" one of the first things I get as a response is "oh you've probably been reading those alt-right sites" or some variation of it. Most articles with controversy make sure to have a demonized group, so that people won't start getting odd ideas like going against the narrative because its bigoted and or racist.

Social Justice as a marketing tool.
 

Lou Wrong

Butthurt about Buddyloids
kiwifarms.net
The funny thing is that Web 2.0 was a lie. Web 1.0 already had interaction, whether it was fire blowing out of your cursor or signing a fucking guestbook and seeing a visitor counter. Web 2.0 is like Fifa 2019 ; a marketed name for what was a tiny gradual change.
If anything, the opposite has occurred. Never has the average user's opinion mattered less than it has at present and its all due to the involvement of "social media" companies and the filtering that they use.
 

Medicated

Not the fun kind
kiwifarms.net
If anything, the opposite has occurred. Never has the average user's opinion mattered less than it has at present and its all due to the involvement of "social media" companies and the filtering that they use.
Yeah we already know Google filters results after the lawsuit they recieved. If thats what we know they are filtering, what else is there? Not to mention various major websites manipulating trending results to steer conversation.
 

Jaimas

YOU AND ALL YOUR KIND ARE COWARDS!!
True & Honest Fan
kiwifarms.net
It's completely a problem of platforms that are too big for meaningful moderation. Things like facebook or youtube. Small communities united by a common interest don't devolve in to "toxic" behaviour. This forum is a case in point. Large enough to have off-topic discussion with people who have varied opinions, but it never or at least rarely devolves in to people going full blown autism on each other. There's moderators who do janitorial tasks, but it's not a full time job.

This wouldn't be a problem at all if only big companies like Google and Facebook would stop trying to capture the entire internet in their own silos. They have all these people with no uniting common interest all chimping at each other like battery hens and they have to invent fucking AI to try to sanitise it enough for advertisers not to throw up their hands in horror.
You say that like it was a bug and not a feature. They pushed AI and algorithmic tools because that was the future they wanted.
 

Maggots on a Train v2

new and improved account
kiwifarms.net
If anything, the opposite has occurred. Never has the average user's opinion mattered less than it has at present and its all due to the involvement of "social media" companies and the filtering that they use.
The average user's potential reach got min-maxed. Back then, a few hundred or thousand people talked to each other. Now, you get ignored if you aren't the rare person who can update regularly and be a micro-celeb.

You can't even have real conversations on most big sites, because they are deliberately designed to constantly cycle through content and bury it.
 
Tags
None

About Us

The Kiwi Farms is about eccentric individuals and communities on the Internet. We call them lolcows because they can be milked for amusement or laughs. Our community is bizarrely diverse and spectators are encouraged to join the discussion.

We do not place intrusive ads, host malware, sell data, or run crypto miners with your browser. If you experience these things, you have a virus. If your malware system says otherwise, it is faulty.

Supporting the Forum

How to Help

The Kiwi Farms is constantly attacked by insane people and very expensive to run. It would not be here without community support.

BTC: 1DgS5RfHw7xA82Yxa5BtgZL65ngwSk6bmm
ETH: 0xc1071c60Ae27C8CC3c834E11289205f8F9C78CA5
BAT: 0xc1071c60Ae27C8CC3c834E11289205f8F9C78CA5
LTC: LSZsFCLUreXAZ9oyc9JRUiRwbhkLCsFi4q
XMR: 438fUMciiahbYemDyww6afT1atgqK3tSTX25SEmYknpmenTR6wvXDMeco1ThX2E8gBQgm9eKd1KAtEQvKzNMFrmjJJpiino