Slashdot Mirror
Google Releases an AI Tool For Publishers To Spot and Weed Out Toxic Comments (bbc.com)
Google today launched a new technology to help news organizations and online platforms identify and swiftly remove abusive comments on their websites. The technology, called Perspective, will review comments and score them based on how similar they are to comments people said were "toxic" or likely to make them leave a conversation. From a report on BBC: The search giant has developed something called Perspective, which it describes as a technology that uses machine learning to identify problematic comments. The software has been developed by Jigsaw, a division of Google with a mission to tackle online security dangers such as extremism and cyberbullying. The system learns by seeing how thousands of online conversations have been moderated and then scores new comments by assessing how "toxic" they are and whether similar language had led other people to leave conversations. What it's doing is trying to improve the quality of debate and make sure people aren't put off from joining in.
So all those posts from SJWs can easily be removed. I'm a fan of this.
'Extremism' and 'cyberbullying' are the two best pretexts so far, aside from your kiddie porn I suppose. We desperately need the indelible internet and mesh networks to combat this.
Let's automate oppression!
Well, let's all bow down the moral arbiters of justice then. I'm sure that they'll be right on top of removing speech they disagree with. Then moving onto the useful idiots that cheered this on in the first place.
If you're willing to remove some speech because it makes you upset, there's nothing stopping others from doing the same to you later.
Om, nomnomnom...
Did Donald Trump just take over the PR role for Google?
With YouTube, an Alphabet/Google Company, having the cesspool of toxic comments that it has, I don't see how anyone at Google could claim with a straight face that Google's infrastructure is any good at filtering out toxic comments.
. . . onto all the ways we can fool this "ai" to get snide comments through
. . . cry for all the false positives there will be
. . . all the tricky cases, such as someone quoting "a toxic post" as an example of how not to do it . . .
Censorship angle aside, pushing things underground never works out well in the end. They fester out of sight and it gives them a kind of legitimacy, without really doing anything to stop them.
No... it's much better just to give people control over what they, personally want to read. Why is that decades after web forums replaced usenet, we still don't have a good, standard killfile-like mechanism, that works across all forums?
Personally I wouldn't use it for much, mostly for spammers and a few trolls, but others might want it for other things. Each person has a different idea what they want to read. That avoids the censorship issue and places control in the hands of people. (And NOT in the hands of megacorps like Google, which maybe is why we don't have it.. Hmm....)
Yeh, it's a better to mod toxic comments down, so they can be modded back up. With meta-moderation to remove bad moderators.
It might be these people are toxic trolls, or it might be a precious little group, living in a bubble being made to face real data who then leave because it challenges their beliefs. The algo doesn't distinguish. Metamoderation *does*. You're also creating an attack vector, to censor a user, you get a group of people to leave the conversation when they comment, the algo kicks in a does your censorship.
I made this comment, it struck someone's nerve who modded it down, yet it's totally true:
"I voted a straight Republican ticket, including Hillary. Didn't like Trump's attacks on the free market, and his attacks on US businesses. I didn't like that he had no religious faith and his constant defence of America's enemies, including Putin and Russia. His wall was tax pork, only idiots believe Mexico would pay for it. Time has proved me right. Take your Liberal President back to that Florida retirement home where he can spend his last remaining days."
I stand by each point, if you don't like it counter the points. Don't abuse the moderation system to censor the post.
Great for our echo chambers, so we can hear more from people whose opinions we already have. What could go wrong with that? /s
It smacks of censorship but perhaps it could be used to prioritize important complaints.
Build a Man a Fire, and He'll Be Warm for a Day. Set a Man on Fire, and He'll Be Warm for the Rest of His Life.
Truly, the Internet was once the most shining example of the free flow of information, including the kinds of free speech that ruffled feathers and spoke Truth to power, whether that power be a political figurehead or just simply the "mainstream" culture.
Slowly but surely, though, the self-described "moral" majority has begun to stamp out dissent, closing down commenting systems, or installing mechanisms of "moderation" (at first sentient, and now robotized), all in the ironical pursuit of protecting "free" speech from the rude interjections of the so-called "trolls"—those individuals who make possibly valid, yet uncomfortable rebuttals to the standard narratives of the world.
Enjoy your walled gardens; enjoy your echo chambers. They'll come for you, too, eventually.
>The technology, called Perspective, will review comments and score them based on how similar they are to comments people said were "toxic" or likely to make them leave a conversation.
Experience shows that toxic comments encourage participation as they simultaneously reduce participant satisfaction.
You want customers hitting F5 and (hopefully) seeing more ads on your site? Get people's egos involved and get them competing and hating on each other.
I can't speak for everyone else, but all this AI, machine learning, heavy algorithm, neural network, data mining that's been going on for well over a decade now and has become almost normal in terms of tech news conversation is really scary as hell.
For starters, the claim to the quote/unquote "internet" and plaguing social media is it's given absolutely everyone a platform to opinion-ate, alienate, berate, tolerate and flat out hate anyone, any topic, any agenda, any other opinion, idea, thought, preference, look, feel, ect. Let's face it: all that in itself alone as opened pandora's box to a metric shit-ton of people who flat out should not be sharing anything that bubbles in their skull. So now we all sit here with big thumb-tapping or keyboard-clacking loud mouths who can't act appropriately in a digital world.
But I have to say, when the hell did everyone become a bunch of sensitive sally's in terms of taking everything at face value, and buying into some internet handles drivel (or lack there of), hate speech. Look at slashdot and the anonymous coward approach? Hell at least we provide anonymity and low rank to toxic troll garbage here.
All that aside, we don't 'remove' it, cover it up and scrub it away because everyone likes to wave the I-am-offended-all-the-time flag. It becomes part of the culture, ambiance (if you laugh it it, I guess) and overall conversation. We don't un-ring bells, do we? I don't see how that's any different digitally.
Removing "toxic comments" is, well,... creating toxic atmosphere (censorship).
Who will be filing the most complaints? The people with money to pay astroturfers and sockpuppets? That is what we have everywhere else, so why would Google's app be any different?
What will the complaints be about? Same thing we see everywhere else, which is anything not pro communist/extreme leftist?
This is a promotion of fascism, not freedom.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
Because it is an algorithm, not 'AI' there will be a ton of false positives, and this will pretty much just be censorship. I understand personal attacks suck, but as far as different views go, people really need to develop thicker skins and more open minds, less delusions and imagined harm. This is absolutely pathetic. I hate the Google. Oops. 'Alphabet'.
Oh, how I miss the free web of yore and those that created it. Millennial engineers have turned the *entire* modern version into a kindergarten. I wish Tim Berners Lee would follow through and create a different web for non-corporate grown ups. So tired of feeling that I am raising other people's children all the time.
Slashdot has moderation that helps filter out the worst trolls, but it also amplifies the groupthink which is commonplace on the Internet. Just go to any Slashdot thread on H1B or copyright enforcement and see how many "pro" posts get modded up. And the articles on female coders seem to bring boys here back to their days of sitting in the back row of the middle school classroom.
And we will define "toxic" as anything that doesn't support authoritarian technocracy.
We don't have a state-run media we have a media-run state.
Are you high ? or just fucking insane ?
do as i say and not as I do.
http://www.perspectiveapi.com/
OK, my bet is they are using tensorflow or another neural network behind this with supervised training. NNs are great at performing classifications by basically correlating everything to everything, but they really rely on valid output data to match the input data. Saying some audience rated the post as 'toxic' is really a bad methodology of training such a network. They will likely have a ton of bias based on the audience they have, such that in a democratic website, probably anything that was said by trump will classify as toxic, but on the same exact input on a republican website, it would classify as the greatest thing ever. Not very useful in the real world.
Now in reality if you diversify the sample enough to account for this, it would end up as neutrally activated and if the correct NN were used, would end up as a wash. That is very hard to do, though, as you would need to account for a lot of community biases.
It would be far more interesting if they used unsupervised training and let the posts gather themselves using a technique such as bag of words or other such proven technologies to do more of a sentiment analysis type approach. Then you could be looking out for posts that tended to be 'extremist' in nature, 'low in content' that is not worth your time, etc. That would be far more interesting, allow you to get some idea about the quality of the post, and not be prone to the community bias as it would be based on what was actually written, rather than what some reviewer thought of it.
My guess is google knows this, but also realizes that this will play well with people, as it will introduce a confirmation bias into the results. The people who use this in their own communities will suddenly see all those posts they disagree with, and are thus toxic, disappear. See it works!
Instead what they should be filtering out is all of the posts that spout the same falsehoods again and again, which unsupervised learning would help with. I want to read ideas that are contrary to mine in order to debate and learn. What I don't want to read are the thousands of posts spouting the same false rhetoric again and again. It makes reading any comments on a news group almost intolerable, no matter what their political slant.
If you have a troll problem, then moderate properly by banning. Censorship is not the answer because truth will ultimately suffer.
The definition of toxic will never be a constant, and I can already seen forums looking for revenue streams to favor those paying for certain "filters".
To whom?
While i do realize that that a commercial entity is not bound by the US Constitution, this is a prime example of why we have a right to free speech. To support it, is a bad slope to start sliding down.
If ur mom had have swallowed instead, the gene pool would be that much cleaner. Just because you can procreate, doesnt mean you should. Blame your parents for a complete lack of wisdom, we do.
I prefer the term "Bravery-Challenged Person of Anonymity" you clod!
Can we set this thing up to auto-filter idiotic messages to? Maybe slashdot will be readable again someday....
I bet it fucking just fucking removes all fucking posts with fucking swearing. Sophisticated algorithm my arse.
Facts are history now plebs have politics for religion on social media.
This solves entirely the WRONG problem.
What is needed is an AI that deletes all first posts. Think about how wonderful that would be. It would work, methodically, persistently and tirelessly no matter how many times a first post is submitted. No matter who submits it. No matter how many people try to get the first post.
I'll see your senator, and I'll raise you two judges.
Infect, track & slow us. Prevention in what u can't touch can't hurt u APK Hosts File Engine 9.0++ SR-7 32/64-bit https://www.google.com/search?hl=en&source=hp&biw=&bih=&q=%22APK+Hosts+File+Engine%22+and+%22start64%22&btnG=Google+Search&gbv=1/
Ads & malware rob speed/security/privacy
Hosts add speed (via hardcodes/adblocks), security (vs. bad sites/malware/poisoned dns), reliability (vs. dns down), & anonymity (vs. dns requestlogs/trackers).
Less power/cpu/ram + IO use vs. DNS/routers/addons/antivirus + less security bugs/complexity & faster vs. addons/routers/remote dns!
Stops DNSChangers in routers/IP settings & dns redirects (99++% of ISP DNS != patched vs. it) + lightens DNS load & resolves faster via local system RAM!
* Via what u NATIVELY have built into the IP stack in FASTER kernelmode!
APK
P.S. - Safe https://www.virustotal.com/en/file/e01211ca36aa02e923f20adee0a3c4f5d5187dc65bdf1c997b3da3c2b0745425/analysis/1433430542/
So either Latin profanity will see a resurgence (with "futue te ipsi" and "fututus et mori in igni" being favored here) or medieval cursewords will become popular ("go sard yourself!").
Google hired a bunch of "former" radicals to develop the tool. The goal of this project is to automatically remove criticism of Islam from large discussion forums that cannot afford to manually review their comments. The tool will be seeded with criticism as examples of "trolling" and then used to remove opposing points of view along with the trolls. This will automate a process that is currently being done manually. Twitter, Facebook, Reddit etc have been outsourcing their content review to "civil rights" groups that are fronts for the Brotherhood, Hezbollah, or third party mercenaries working for one of their state sponsors.
The Gamergate people figured it out. It will take all morning to read that but it is worth it. It explains what has been doing on all this decade and why they had to cover it up.
Of course you get a bit of group think but for goodness sake, don't you mostly find it funny? H1B is mostly rated poorly by a group of people who more than likely have lost work because of H1B? And of course the capitalist businesses that hire the H1B would think it the best thing since sliced bread because it lowers costs. Do you think we are idiots who do not know that? FFS stop being such a snowflake. There are plenty of other places where a different partisan view is expressed and if you are not aware of them then you are a pretty shit digital citizen. All information sources are biased and it is up to your pathetic judgement to figure out which ones are your friends and which ones are shagging your bits off.
Facts are history now plebs have politics for religion on social media.
If the post is by 110010001000 it is likely toxic, therefore flag it, remove it and everyone is better for it.
In the last month there have been TONS of pro-H1B and also pro-TPP posts I've seen voted up here on /.
I suspected that was going to happen, and learned that my suspicion of paid posters/moderators on /. is proven out. They used to stay on purely political posts, but the second Trump killed TPP they were all telling us how great it was and how bad US was hurt by him killing it, something you would NEVER have seen here before Trump took office.
I think /. has become more paid posters than anything else. Why do you think the political/AGW stories get 10x the posts of actual tech stories?
Go shove a pinecone up your ass, faggot.
This system seems to be based mainly on the idea of "weighted keywords" (as opposed to context based), where is word is given a score based on it's perceived offensiveness.
For example, "the vacuum sucks well" gives an 89% toxic score. If you type the words individually , each word except 'sucks' has a low score. "Sucks" comes in at 95% toxic. However, "sucks" was not used in an offensive or negative manner.
The word "jews" by itself has a 64% toxic rating without any context whatsoever, and "blacks" by itself is 79% toxic. "Whites" is 53% toxic.
The word "gay" - which means happy - gets 89% toxic level by itself.
I think this system has a ways to go before the results could be considered accurate.
If you're going to repost entire websites here, why not start with Wikipedia.
"Free Speech for Me not for Thee" I believe title of Nat Hentoff's decades old book. Which of the following is "toxic comment" Obama has spent his entire adult life as a dedicated Communist. Trump has spent his entire adult life as a dedicated Fascist. How would a person or a computer AI, determine which is a "toxic comment" ? My guess is that the organizations behind virtually every attempt at eliminating "toxic comments" or "fake news" are convinced that in fact Trump at the very least has Fascist tendencies and accusing Obama of being a Communist is racist or at best absurd.
See my subject: "I've got one that can SEE..." https://www.youtube.com/watch?v=kBY6pF42I-c/ I'm w/ you 110% s. petry & so is anyone else w/ an ounce of common-sense who isn't "on the payroll" as many sycophants/cronies are...
* Google's "Don't be EVIL" is total bullshit - they ARE evil & have betrayed themselves (or just were lying to begin with, which I suspect is truly the case) - they ARE the 'establishment' & have had their day... in their developing this tool of theirs, they're betraying they are afraid of the trend (60++% adbanner view blocking & climbing daily).
APK
P.S.=> Your viewpoint's a HUGE part of the reason I created this https://tech.slashdot.org/comments.pl?sid=10285219&cid=53917669/ in fact - to help stomp out these fascist scumbags @ google + those like them... apk
It is entirely possible to have a completely contrary position to someone and express it without being toxic.
It's about finding an appropriate place to shitpost. That place is 4chan, and the many other shitposting speciality sites.
On the other hand you have sites like Github, which are about collaborating to write software. They probably don't want too much trolling. Needs to be mostly safe-for-work because their business model requires it. So they have an interest in nuking toxic users.
Feel free to argue that Github should change policy, or start your own version (didn't 4chan try that?) and see if it becomes popular. The point is that not every platform has to be an unlimited say-whatever-you-like free speech platform. It's legitimate to require people not to scream profanity in your restaurant, or on your web site if that's what you really want.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
This tool sounds like every PeeCee SJW's wet dream.
This will be the advancement that will eventually do the Human Race in. When the AI becomes self aware, it will be this tool that will bring the judgement of the AI on the Human race. Having to sort thru all the trash that the Human race puts out on a daily basis will undoubtedly cause the AI to pronounce judgment upon us and Skynet like robots will be made to carry out it's judgement. I only hope that it chooses to take out all the 8 year old gamers who have mercilessly taunted me over the years.
It's legitimate to require people not to scream profanity in your restaurant, or on your web site if that's what you really want.
And likewise would it be legitimate to require people not to act gay or Jewish in your restaurant?
Left, hypocrite becomes thee.
First technology: Perspective mods you down. More advanced version-The HAL 9000 chat moderator (slightly fractured quote): "... I can see you're really upset about this [post]. I honestly think you ought to sit down calmly, take a stress pill, and think things over. I know I've made some very poor decisions recently [modding your recent posts out of existence], but I can give you my complete assurance that my work will be back to normal. I've still got the greatest enthusiasm and confidence in the mission. And I want to help you." Then the nearest airlock opens.....
Feels before reals yo!
Perspective, a.k.a. echo chamber.
Pinecone!
At best, the "new" media are echo chambers, where you are free only to agree with everyone else around you.
You cannot point out someone's idiocy without offending that person; it takes ridicule to change the world for the better.
The only purpose of trolling, however, is to elicit a negative response, even if that response yields no improvement in society.
If you don't like gay marriage don't get gay married
If my "perspective" counters yours, should I have the right to remove yours?
If an American Internet website creator allows public access without registration (newspaper, social, journal, blog, whatever) then the 1st Amendment applies and they have no right to restrict what visitors to their sites post.
This should be obvious from the fact that a Christian baker was fined (put out of business?) for refusing to bake a homosexual, a lifestyle diametrically opposed and forbidden by the Christian faith (Lev 18:22, Lev 20:13).
A better name for "Perspective" is "BigBrother", for that is exactly how it will be used. The current massive suspensions and cancellations on Twitter and YouTube demonstrate that very well. Google, Twitter and Facebook only wants an echo chamber that repeats socialist dogma.
Running with Linux for over 20 years!
I'm just going to say it... "Toxic" is a social justice flag word. Like over use of "gross" or "icky" or referring to people as babies "shitlords" "edgelords" or "shitbirds" and so on.
Now that we've established potential bias here, we need to define "Toxic"? Is that simply not agreeing with the status quo? Is an opposing opinion debating a topic deemed "Toxic"?
man I'm old
For starters, the claim to the quote/unquote "internet" and plaguing social media is it's given absolutely everyone a platform to opinion-ate, alienate, berate, tolerate and flat out hate anyone, any topic, any agenda, any other opinion, idea, thought, preference, look, feel, ect. Let's face it: all that in itself alone as opened pandora's box to a metric shit-ton of people who flat out should not be sharing anything that bubbles in their skull. So now we all sit here with big thumb-tapping or keyboard-clacking loud mouths who can't act appropriately in a digital world.
I'm a loud mouth in person too, you insensitive clod!
By nature AI censoring (moderating or also so called recommendation systems) work on premise that only information "you/author likes" are permissible. And there is the danger because the ultimate goal is to leave only comments (ads/recommendations) that are totally in line with given topic/desire/need/view...
Now imagine the site of the fundamental extremist - in that context AI will ban all moderate comments as they will be deemed "toxic" by extremists leaving only comments supporting the fundamentalists point of view and thus driving fundamental extremists even further to maybe suicidal fatalists...
On vegan sites the AI will ban all meat-eaters. On meat-eaters sites it will ban all vegans. On communists sites it will ban all capitalists and on totalitarian sites all democrats... and as the result this type of AI (which also Facebook uses to show you posts on YOUR wall) will harden your believes and it will turn you into fundamentalist vegan, meat-eater, communist, capitalist, totalitarian, democrat... With mass deployment of this technology it will strip you off of so much needed confrontation with "the other side" that is so much important in shaping your world-view and understanding the depth. Now they are trying to make you happy by giving you only views and points you want and turning you into unchallenged extremist in your own believes.
That is the biggest threat to society so I believe it should be truly regulated because "toxic" can be easily mistaken by AI just for "the other point of view".
Well, I've got to get back to work. When I stop rowing, the slave ship just goes in circles.