Google Says AI Better Than Humans At Scrubbing Extremist YouTube Content (theguardian.com) 136
An anonymous reader quotes a report from The Guardian: Google has pledged to continue developing advanced programs using machine learning to combat the rise of extremist content, after it found that it was both faster and more accurate than humans in scrubbing illicit content from YouTube. The company is using machine learning along with human reviewers as part of a mutli-pronged approach to tackle the spread of extremist and controversial videos across YouTube, which also includes tougher standards for videos and the recruitment of more experts to flag content in need of review. A YouTube spokesperson said: "While these tools aren't perfect, and aren't right for every setting, in many cases our systems have proven more accurate than humans at flagging videos that need to be removed. Our initial use of machine learning has more than doubled both the number of videos we've removed for violent extremism, as well as the rate at which we've taken this kind of content down. Over 75% of the videos we've removed for violent extremism over the past month were taken down before receiving a single human flag."
Didn't we already have a post about training AI's? (Score:4, Insightful)
Who is training the AI and deciding what is illicit? Better by what measure?
Re: (Score:1)
I was wondering something similar.
Does the AI also flag our (err... the US') domestic religious extremism?
Re: (Score:3, Insightful)
As long as it's not muslim of course. That would be 'racist' and 'islamophobic.'
I also wonder if it allows 'extremist' content in the same vein as Galileo's observations about the solar system. Somehow I think this will end up as a simple anti popularity filter.
Re: (Score:2, Insightful)
Not really applicable, because Galileo's science had nothing to do with him "getting into trouble". He had direct permission from the Pope to publish the science, and it was hardly "shocking", since the model was well known since Copernicus.
And, incidentally, we now know neither heliocentrism nor geocentrism is scientifically "correct", since Einstein it's well established that reference frames are arbitrary, and the only scientific difference is the complexity of the drawing needed to equally-correctly de
Re: (Score:3)
And, incidentally, we now know neither heliocentrism nor geocentrism is scientifically "correct", since Einstein it's well established that reference frames are arbitrary, and the only scientific difference is the complexity of the drawing needed to equally-correctly describe the orbits.
That's not how relativity works. Yes, you get to pick any reference frame you like, but the Earth still revolves around the Sun.
Re: (Score:1)
Only in the sense that's the simplest human conceptualization.
Gravity acts on all the planets and the sun, collectively, resulting in particular paths. "Around the sun", when the sun is just another object also being acted upon by gravity from within and without the solar system, is simply a more convenient model. It is not "more true".
It is the methodologically best model to use per Occam's Razor, but contrary to further misapplication of that particular theist's concept, the Razor speaks to efficient co
Re: (Score:2)
Relativity applies to inertial frames of reference. Gravity exerts an acceleration, so a planet orbiting the sun is not traveling in an inertial frame.
Re: (Score:2)
And, incidentally, we now know neither heliocentrism nor geocentrism is scientifically "correct"
The earth orbits the fucking sun. Not the other way around. Einstein would happily correct you on that if he could.
Re: (Score:2)
Re: (Score:2)
i thought Galileo got into trouble because he didn't know when to shut up. Publishing his knowledge was allowed, just, but for the rest he had to keep a low profile so as not to threaten authority. He wasn't good at keeping low profile.
Also I thought Galileo introduced the idea of relativity of reference frame, Not Einstein. Then Newton included it and extended the laws of movement.
Then it got into trouble late 19th century and Einstein reworked it/saved it so it could take in account electromagnetism. Poin
Re: (Score:1)
Re: (Score:3)
Modded down for being pedantic, condescending, and completely missing the fucking point.
Are you talking about Galileo now?
Re: (Score:2)
Google nuked Dr. Jordan Peterson (not only his Youtube channel but also his GMail account). Restored after an uproar, but I think that tells us all we need to know about their new AI-in-training.
Re:Didn't we already have a post about training AI (Score:4, Funny)
I believe the training set is curated by a group headed by Coraline Ada, Christopher Poole, Deray McKesson, and Linda Sarsour.
Re: Didn't we already have a post about training A (Score:1)
The ADL, of course. Can't have the goyim knowing. ;)
Re:Didn't we already have a post about training AI (Score:5, Informative)
4chan is training their AI. People on the /pol board have been running a mass false-flagging campaign lately, getting videos and channels they don't like taken down. When the victims appeal or complain they just get a robot "reviewing" the strike and ultimately have to resort to tweeting at YouTube staff to have any chance of getting things resolved.
YouTube's content filtering is out of control, anti-free speech, kafkaesque and devoid of any human oversight to catch errors.
Re: (Score:1)
Re: (Score:2)
Who is training the AI and deciding what is illicit? Better by what measure?
That's the first thing I thought. Who or what is judging whether humans or AI are better at judging - humans or AI ?
Re: (Score:1)
"Better" means that the test set had a better precision and recall than when the humans processed it.
This is, of course, in theory. In practice, it is likely that the training / test sets were so similar that it will not scale in the real world.
Too little too late (Score:5, Insightful)
Sad thing is I'm gonna miss the ad supported internet. I'm pretty good about ignoring ads (and will cheerfully click on any ad that annoys me to give 'em a false positive) so it never bothered me.
Re:Too little too late (Score:5, Insightful)
Advertisers could have saved themselves by agreeing to a non-offending standard (no popups, no autoplay videos, etc.) for advertising that would then be accepted by ad blockers. But no - now their business model is vanishing.
Those ads are completely ineffective (Score:2)
Re: (Score:2)
What I have in mind is an industry-developed standard that would enforce its branding on all advertisers. Ad blockers would have a Settings checkbox for accepting ads that comply with the standard. You would be motivated to use the checkbox as an alternative to having to disable your ad blocker for every single news site you encounter in a search.
A "bad actor breaking ranks" would be subject by the standard consortium to the same set of weapons-grade legal attacks that apply to content piracy. More importan
Re: (Score:2)
AI == Unaccountable (Score:1, Informative)
Pewdiepie is alt-right? The advertisers pulled their ads because they were shamed into doing it by a smear campaign orchestrated against centrists by authoritarian, puritanical leftist bullies.
Re: (Score:1)
Once they figure out how to serve ads that aren't a massive attack vector, I will consider this.
Re: (Score:2)
Sad thing is I'm gonna miss the ad supported internet.
No. I used to think that, but it's pushed the average quality of the internet down. Clickbait is too easy to find, good stuff too hard.
Maybe (Score:5, Insightful)
Did the AI get programmed to only target politically incorrect content? Unfortunately double standards for YouTube, Twitter and other sites are all too common. Color me skeptical that this is anything other than an automated political censorship tool.
Re: (Score:2)
At last, a high-paying tech job for that cousin with the transgender theory degree who didn't manage to pass her LSAT.
Re: (Score:2)
Re:Maybe (Score:4)
It's more like just "political content". Lots of people on the left are getting their videos taken down too, mostly due to people maliciously flagging them. Shaun & Jen, H. Bomberguy, Contrapoints... Some people have lost entire channels.
https://youtu.be/UgNhO8lMINw [youtu.be]
Re: (Score:3)
"Controversial" = "Not SJW" (Score:1, Insightful)
BLM video spewing anti-cop or anti-white hatred? That will be flagged "okay."
Conservative video suggesting MAYBE Trump is right about immigration? Oh, that's HATE speech--permanent ban!
Re: (Score:3, Insightful)
You know, for a political party that complains about their opposition being a bunch of "snowflakes" that need special treatment, you guys do whine a LOT about how you aren't getting special treatment.
Re: (Score:1)
No special treatment? What planet do you live on? They get so much special treatment it's scary.
Of course it's all bad. So many Internet corporations, media outlets and NGOs have gone censorship crazy ever since the conservatives won the election. Accounts getting suspended, videos/posts getting censored, media attack campaigns, etc are all special treatments.
Re: (Score:2)
What planet do you live on? They get so much special treatment it's scary.
I took the original post to say not that they weren't getting special treatment (which they do), but that they were complaining they weren't getting more of it.
Comment removed (Score:4, Insightful)
Re: "Controversial" = "Not SJW" (Score:1)
Re: (Score:2)
Re: "Controversial" = "Not SJW" (Score:2, Insightful)
I'm not alt right. I'm a liberal.
Diversity is a code word for less whites.
There is a demonstrable attack on white people, particularly men, in today's society. Your response is an example of that. Someone asks simply to not be censored, and you scream them down with a torrent of profanity and insults.
You are the reason why we have Trump as a president. Stop with the identity politics, the - isms, and the hypocrisy. If you truly believe in equality and the destruction of racism, stop being a fucking racist.
Re: (Score:1)
I don't agree with calling that an "attack on white people" but I fully agree that what the US calls "liberal" or "left" is obsessed with identity politics. In say France, identity politics is mainly a tool of a few small extreme right groups though that may evolve in the near future.
In the US it's useful as a diversion, basically just so that campaign propaganda can say "vote for me because I'm a wimmin".
It's also useful because the US "left" stands for neoliberalism, i.e. a totalitarian laissez-faire and
Re: (Score:2)
Grow a fucking spine. You sound like a fucking whiny 2-year old toddler: "There are other people out there! No! Cuddle meeeeeeeee!"
Re: (Score:2)
Re: (Score:2)
Re:"Controversial" = "Not SJW" (Score:4)
Plenty of progressive/left leaning videos get taken down too. 4chan users have been flagging them like mad lately, and some videos and even entire channels have been removed because of it.
It's not some political conspiracy by YouTube, it's just a shitty system that is wide open to abuse and utterly devoid of human oversight.
Re: (Score:2)
How is accuracy measured? (Score:5, Insightful)
The real answer is that there is some human who decided whether a video is extremist, then gave it to both the AI and another human, and the AI was able to learn how to agree better with the 1st human. Unfortunately, that doesn't actually tell us if the video is extremist or if the AI is any good, because the first human isn't more right than the second.
Re: (Score:2)
Re: (Score:2)
Re:How is accuracy measured? (Score:4, Interesting)
I would imagine that Google has some content policy experts who are making up the rules of what's acceptable and not. And then a whole lot of whack-a-mole workers trying to police by those rules. I guess that doesn't say if the expert is right, but the AI is doing the leg work more consistently in line with policy than the people. If you let say 5% through to a second opinion by humans or if there's some sort of appeal process there should be a pretty continuous learning. Though I suspect the AI will completely fail to understand parody or other use of extremist elements in a non-extremist way.
Re: (Score:2)
Translation: Google is working really hard to boost alternative video networks.
Re: (Score:2)
As a Twitter alternative I use Gab.ai, which is about as free as a wide-open-to-all social network gets. The sort of place where if you don't like someone's opinions, go post your own, anything that's not doxxing, direct threats, or outright illegal.
Alt-video is harder, if only because it's relatively expensive for storage and bandwidth. Notions have come up about distributed content, but I think that might get further with micropayments for hosting it.
Re: (Score:2)
Let's be fair, even some humans completely fail to understand parody.
Yes, frequently even the people who claim to be creating parody.
Re: (Score:2)
This. I'm thinking in particular of Hellsing Ultimate Abridged, episode six, which not only features Nazi Vampires, but an extended rant from Bishop Maxwell to a horde of hood-wearing Papal Knights that begins with "F*CK THE NEW POPE!" It's satire, but does the AI allow for that?
Re: (Score:2)
You're right. The issue is how do we get a more accurate test set to evaluate the algorithms. Typically the test set has the same biases and accuracy a
Re: (Score:2)
If all you have is no brain, everything looks like a slippery slope.
Re: (Score:2)
http://edition.cnn.com/2017/01... [cnn.com]
"... For Blasphemous Facebook Comments In Pakistan" (June 13, 2017)
https://tech.slashdot.org/stor... [slashdot.org]
Social media is getting interesting around the world. The SJW are busy banning history.
Re: (Score:1)
I live in Canada. There was M-103 in March and C-16 in June. And those are only two minor examples of the ongoing obliteration of free speech in the Western world. "Hate speech" laws are getting broader and broader everywhere in the West. Words like racism, sexism and gender are also getting new definitions in order to promote a political agenda. We are heading toward a dystopia a la 1984 and it's not a slippery slope fallacy.
Re: (Score:1)
Soon you won't even know what you don't know (Score:2)
Just like the good old days.
I'd be eager (Score:3)
To see what other countries decide is extremist, and needs to be scrubbed. Tiananmen Square videos aren't looked on very kindly in China for example. Or women in bikini's in some other countries.
Does each country have it's own set of what is extremist, or do they all contribute to a global set of extremism, until al that is left is sponsored content from paying advertisers?
Re: (Score:3)
"Does each country have it's own set of what is extremist, or do they all contribute to a global set of extremism, until al that is left is sponsored content from paying advertisers?"
Welcome to the World State in our Brave New World. Our AI will ensure a harmonic and conflictless society without the tensions of previous barbaric societies.
I propose a "law"... (Score:2)
Any AI that can search for unacceptable values of extremism, however defined, in political discourse is one that would be qualified to run for office.
The Key Words are Scrubbing/Remove/Combat (Score:2)
The whole idea of protecting us from bad content stinks. Actually it's outright alarming. Firstly, do we need to be protect the pedophile terrorists that are served as prime safe example of the censorship Google/Facebook/Youtube and others are performing? Secondly it's a very bad idea to protect us from what designated enemies like Russia want us to know, with the policies to eliminate 'fake news'.
Thirdly it's a bad idea to protect us from our own progressive and leftwing activism. A socialist site checked
Re: (Score:2)
And one can guess what will follow. Google is already constraining sites that publish their ads for revenue. Sites involved in whatever activism they're doing regularly get into conflict with Google who withdraws ads. After correcting the site or insisting long enough Google reverts the decision. There still is the option to not use the ads and forego the revenue but the bulk of the people quickly learn to stay out of trouble. I expect Google ads and search results to interact in the future. After all anyon
Re: (Score:2)
I find that more alarming than if Google were only censoring rightwing sites. At least if it were completely skewed, we could tell what's missing and go elsewhere. But when it's all shaved down across the board? then you don't have any point of reference that makes the censorship evident.
What people say or do, outside of the blatantly illegal, should be none of Google's business. Either index it all the same and let ME decide what I want to search for and view, or admit that you're shaping minds and traffic
Jordan B. Peterson (Score:2)
So that is why they shut down Jordan B. Peterson? :D :D
https://twitter.com/jordanbpet... [twitter.com]
Re: (Score:2)
Really Bad Idea (Score:2)
Re: you FAIL it (Score:2)
Well, not proverbial, but you know what I mean. Or maybe you don't. Still, so long as you are here, even at -1, /. has not fallen.