Yahoo's New Anti-Abuse AI Outperforms Previous AI (wired.co.uk) 119
16.4% of the comments on Yahoo News are "abusive," according to human screeners. Now Yahoo has devised an abuse-detecting algorithm "that can accurately identify whether online comments contain hate speech or not," reports Wired UK:
In 90 per cent of test cases Yahoo's algorithm was able to correctly identify that a comment was abusive... The company used a combination of machine learning and crowdsourced abuse detection to create an algorithm that trawled the comment sections of Yahoo News and Finance to sniff out abuse. As part of its project, Yahoo will be releasing the first publicly available curated database of online hate speech.
The machine-learning algorithm was "trained on a million Yahoo article comments," according to the article, and Slashdot reader AmiMoJo writes "The system could help AIs avoid being tricked into making abusive comments themselves, as Microsoft's Tay twitter bot did earlier this year."
The machine-learning algorithm was "trained on a million Yahoo article comments," according to the article, and Slashdot reader AmiMoJo writes "The system could help AIs avoid being tricked into making abusive comments themselves, as Microsoft's Tay twitter bot did earlier this year."
"Hate speech" (Score:5, Insightful)
Daily reminder that if you subscribe to this idea of "hate speech" you are totally insane and have no mind of your own.
Speech can be hateful, leave it at that, everyone knows it well, no need to make a special idea for it that can also be manipulated to cover valuable critical thought.
Re:"Hate speech" (Score:4, Insightful)
yahoo news and abcnews hate dissenting non-SJW type viewpoints and label them "abuse"
Re:"Hate speech" (Score:5, Insightful)
Re: (Score:1)
It's not like its just Yahoo or a couple of other companies. This is a unilateral trend across all the media.
This is completely obvious, and it should be just as obvious that the kind of people in the general population who support censorship aren't the kind of people who come up with ideas on their own. The common excuse of "that's just what people are interested in" needs to die. Supporters of censorship are being directly manipulated but are given credit for having a respectable point of view. How can yo
Re: (Score:1)
It's not like its just Yahoo or a couple of other companies. This is a unilateral trend across all the media.
You can say that again. The PC culture, or SJW culture, or whatever you want to call it, is probably going to end up destroying itself sooner or later though:
http://www.huffingtonpost.com/... [huffingtonpost.com]
Nobody, not even the BLM movement itself, actually cares about a black life unless it involves a cop, (something Al Sharpton proved) so I think they're probably going to get their wish as soon as the current generation of police officers retire. It will be interesting to watch, from a distance, (we don't have this probl
Re:"Hate speech" (Score:4, Interesting)
I remember in Cincinnati back in the late 90's or early 21st there was a police shooting of some black guy in one of the shitty black neighborhoods. There were the usual protests, calls for the heads of the evil racist cops, etc. The police there responded by simply parking their cars and stopping patrols in said shitty black neighborhood. Crime shot up 600% in the shitty black neighborhood almost immediately. Pretty soon the same people attacking the evil racist cops were literally BEGGING them to come back. Problem solved.
Re: (Score:2)
This is the argument used by dictators everywhere. "Look, I might occasionally murder some people, but if I wasn't around to keep the peace things would be so much worse!" Sometimes they are even right, e.g. in Iraq.
Re: (Score:1)
It's not like its just Yahoo or a couple of other companies. This is a unilateral trend across all the media.
Yeah, and just like I defend your right to say whatever goddamn idiot thing comes to your mind, I also support the right of a company to have policies dictating that they don't have to repeat that idiot thing, just because you feel like they should. Yahoo pays to publish the stuff they want to publish on the internet, and if you want to publish some different stuff? Pay for the hosting! it's pretty simple.
"right to freedom of expression" is not "guaranteed retransmission of your banality." Why is this dif
Re: (Score:1)
Re:"Hate speech" (Score:4, Insightful)
From the article:
The comment dataset came from Yahoo! Finance and News, which you wouldn’t think of as exactly the dank basement of the internet—but it turns out a whopping 7% of comments on Finance and 16.4% on News were deemed abusive by human screeners.
I look forward to perusing this database and finding out exactly how abusive these comments actually are. I don't doubt there are a lot of assholes on the internet. Slashdot has one of the most effective self-policing filtering mechanisms I've seen. Browse at -1 on occasion to see some of the crap, even though I think most obvious trolls have left, as they can't get much traction here. Sure, it goes awry sometimes during heated debates (toss the word SJW in there and you're going to spark some positive and negative mods just with that), but overall does a good job of filtering out most of the garbage. But what's the criteria? Are we talking genuine abuse, or "microagressions", where someone perhaps just expresses an opinion that's not quite PC enough for someone else's delicate ears?
Moreover, a 90% success rate doesn't strike me as all that fantastic. That means 1 in 10 abuse flags is a complete false positive. That would be a completely unacceptable rate with a spam filter. I have a feeling that with all the training that's been done, all you have to do is say one of the "banned" words, like "cunt", and you'll get flagged, even if speaking about it in an appropriate context, like I'm doing. Or am I being abusive just by mentioning the word, since I'm a man? Oops, probably another flag there for mentioning my male gender. Given the tone of my post just now, I'll bet even a human might flag me for a microagression.
Unlike some here, I think the effort to civilize online speech is not unwarranted. Talk to prominent men and (especially) women online and find out how much verbal abuse gets heaped on them. It's fairly disgusting. But this is an area which, I think, needs to be tread upon very carefully. Simply blacklisting hateful speech doesn't cure the problem. It addresses a symptom, and will just push the abuse "underground" to a level where an AI can't detect it. Humans are clever that way.
I'm not sure what can be done short of preventing anonymous interaction, because that seems to bring out the worst in people. Many notorious trolls tend to be cowards, and are mortified when their actions are associated with their real names. This algorithm is obviously one way to address the problem, but I have to admit I remain skeptical about this sort of approach.
Re: (Score:1)
Sure, it's addressing the symptom, which is all that can be done, as there is no cure for assholism. It may be that addressing the symptom is sufficient for the purposes of sites such as Yahoo, since "driving it underground" is effectively equal to "driving it to some other website" and making it someone else's problem.
Re: (Score:2)
I don't know why everyone is so impressed with this algorithm. It's really quite simple:
If poster = heterosexual.white.male Then abuse = true;
Re: (Score:2)
I just can't keep up with all these new computer languages these days.
Re: (Score:3, Insightful)
Exactly. Hate Speech == Censored Speech.
It's fine to have your own opinion but when you dictates that others follow your own myopic viewpoint because you're too insecure, congratulations, you've just resorted to censorship.
Ignoring something doesn't make it go away.
--
SJW, noun, acronym for Stupid Justice Whiner
Re: (Score:1)
Political Correctness
The "Smiley Faced" version of Totalitarianism.
Re: (Score:2)
Sadly, all too true. :-(
Re: (Score:1)
Re:"Hate speech" (Score:4, Insightful)
Re: (Score:2)
On the other hand, only anti-SJWs think that every site owes them a free speech soapbox.
Where do you draw the line? Some material is illegal to host and you will eventually be arrested for distribution if you don't remove it. And between there and Disney there is a huge range, so you are going to have to draw the line somewhere. No matter where you draw it, someone will complain that they are being censored. Yes, there are political parties in Europe who support sexual relations between adults and children,
Re: (Score:2)
Aren't you conflating the two though?
There is a line between "legal content being censored" and "illegal content".
i.e. We're not talking about illegal content, such as illegal numbers [wikipedia.org] (sic.).
We're talking about someone posting their opinions (however unpopular), and others trying to mandat, and dictate that it be removed simply because they are too immature and insecure to handle it.
We already went through this censorship crap with books. As long as someone isn't breaking the law, if you are offended by wh
Re: (Score:2)
I meant things like doxing. I don't know how legal or otherwise doxing is in Twitter's jurisdiction. Revenge porn is another good example.
I have yet to see a single example of someone being banned for posting an unpopular opinion. I keep asking for them, but they are never forthcoming.
Re: (Score:2)
Um... what else do you suggest we call it? (Score:2)
It's the same class as folks who managed to make a concept like social justice into a bad thing. Yeah, there are a few obnoxiou
Re: (Score:2)
>>There are some among us who _want_ the non-stop cavalcade of racist tweets to stop.
Why are you following tweeters who spew cavalcades of racist tweets, or cavalcades of anything you don't like? Un-follow them. Vote them off your personal island, but you shouldn't get to vote them off the planet. Some sick, stupid racist f*ck may want to read what they are writing, and that dumb evil f*ck has the same rights as you.
>>there are a few obnoxious radicals that wanna ruin everybody's fun.
There is
No, there isn't (Score:1)
As for the tweets, I got lazy with my writing and used the word 'tweet' when I should have wrote 'post'. Tweet is the Kleenex of the internet.
And you're missing my point, which is that the left's fringe is fueled by horror stories they see around the world. It's really only been the last few decades that we didn't have regular terrorism practiced against black people and that people didn't look the other way when a husband
Re: (Score:2)
>>the left wing fringe isn't nearly as large as Fox News would have you believe
I don't watch Fox News or listen to talk radio. Conservatives aren't nearly as uniform in their media consumption habits as MSNBC or DemocraticUnderground.com would have you believe.
>>Tweet is the Kleenex of the internet.
Completely disagree. Many, many, many more people post to websites then have even seen a tweet on Twitter. Twitter has a lot of influence in the media and with brand managers, and so second-hand wit
Re: (Score:2)
And you're missing my point, which is that the left's fringe is fueled by horror stories they see around the world.
So what? Censorship does nothing to improve those horror stories, even if by happenstance someone actually involved with one of those horror stories is affected. What makes these would-be censors and their misguided attempts more worthy of our understanding or respect than the people they censor?
If you spend a few years studying that (e.g. if you're rocking an actual history degree as opposed to reading Fark from time to time) then you're gonna be a litter jittery if you're not part of the winning class. In America that's white European males. That's real. That's a thing. It didn't have to be them, but it is. And if you're not one of them... if your on the outside looking in... then you'd be a fool to believe that 5000 years of history was wiped away by 50-100 years of good behavior.
That's not real. That's bullshit. History is not behavior. And 50-100 years of good behavior means everyone responsible for the "history" is dead. Why are you blaming "white European males" (who are those people aga
Re: (Score:2)
You are absolutely right. It is Yahoo's business, and they must do whatever they can to make money with it. And it's not like anything that Yahoo does with its online properties is going to have any effect on the culture. The time for them to have that kind of influence is long past.
But it is very telling -- and sad -- that they are so loudly trumpeting their new language restriction filters. Many boards have such filters (it's not a new idea) but until now no one has bragged about it. And that's the sad
Re: (Score:3, Insightful)
What phrase would you use to describe hateful speech that commercial AIs may wish to avoid using? Keep in mind the target market for these bots is customer service and advertising, so avoiding going full "Tay" with "Hitler did nothing wrong!" is a design goal.
Not everything is a conspiracy against free speech.
Re: (Score:2)
What concerns me most about all this isn't that there is hate speech, or that people find certain things offensive. What concerns me is that more and more, people are no longer being taught to ignore baiting remarks (which is pretty much what anything on Yahoo discussion boards is going to be), but instead to suppress anything that anyone can find offensive.
Does their filtering system allow honest discussion of controversial topics? If it does, then no big deal. But if it simply suppresses any of a particul
Re: (Score:1)
Daily reminder that if you subscribe to this idea of "hate speech" you are totally insane and have no mind of your own.
On the contrary. A person who subscribes to such ideas is a sociopath and is trying to control the minds of others.
Re: (Score:2)
I hate Yahoo's big titted "this will make your jaw drop" headlines in their sidebar "news" stories.
Re: (Score:1)
Re: (Score:1)
Most people are just brainless drones. They aren't really human because they are incapable of critical thought. They aren't really animals either because they're trained to ignore their instincts. They're machines. Artificial intelligence running on fleshy hardware.
"Hate" is a trigger phrase for them. Hearing this causes their programming to load up a particular state of thought and emotion and behavior. They have no will in the matter. They just execute their program and their husk of a body receives pharm
Re: (Score:1)
What is the difference between "hateful speech" (that you say exists) and "hate speech" (that you claim those who believe in are insane)?
tricked (Score:2)
George Carlin (Score:4, Interesting)
Re: (Score:2)
Correct, but not in the way you think. It's the people screaming that everything is PC who are trying to silence others.
It started back in the 80s. People would say unreasonable things like "hay, it would be nice if you didn't name your building after the guy who owned by ancestors" or "just because I'm Asian American doesn't make me your sex fantasy fetish", and the accused would shut them down by pointing out how PC it was.
It's the ultimate defence, because PC is always crazy and never a legitimate critic
Re: (Score:3)
and the accused would shut them down by pointing out how PC it was
Except the accusers were (and are) never shut down by having that observation made. That's the whole point. The rational people point out the nonsense that is the hand-wringing nanny-state-ish PC grievance industrial complex, and the people who are empowered by that movement's nuclear SJW engine are the ones doing the censoring. Yes, PC usually IS crazy, and rarely a legitimate source of critical response, but calling it that does NOT shut down the PC-infected complainer - it just makes them invoke a highe
Re: (Score:1)
Are you trying to *invoke a higher authority*? I ran your post through Google Translate. It's still indecipherable..
Re: (Score:2)
Are you trying to *invoke a higher authority*? I ran your post through Google Translate. It's still indecipherable.
I know, pretending you can't understand things is your hobby. But surely you're not so removed from the actual goings-on in the real world that you understand how in, for example, an academic setting ... a person who cannot tolerate having their world view questioned, and who cannot use the usual PC toolkit to shut up the person questioning their premises and critical thinking skills, reflexively speed-dials the Dean Of Sensitivity and convenes a Social Justice Hearing in order to swing the academic ban ha
Re: (Score:2)
So, what's your specific objection? Is it that you don't find similar behavior in both academia and civil grou
Re: (Score:1)
You always make my day more interesting.
Re: (Score:2)
Re: (Score:1)
What can I say? You're just clowning around, not saying anything outside your own groupthink. You respond the same way to everything, which only reveals a certain pathology, and rendering the content irrelevant. Your state of mind becomes the *subject at hand*. You did the same thing last time, and you're doing it now. That's worth looking into. I, for one, am fascinated by it.
Re: (Score:2)
Re: (Score:1)
...there IS something wrong with what I specifically said
Goes without mentioning... But this is your game, and as they say, the house always wins. However, your projection, though somewhat amusing, is as puerile as it gets, mon petit général
Re: (Score:2)
Re: (Score:1)
You too, don't have *your own words*, as you like to put it, on anything. You just recite mass media propaganda that doesn't interest me the least little bit. But please, do carry on, as you feel you must.
Re: (Score:2)
Re: (Score:1)
Hilarious
Yes it is
Re: (Score:2)
Re: (Score:1)
I am awestruck by the amount of effort you put into this, and very curious. What do you expect to gain?
Re: (Score:2)
Re: (Score:1)
Good morning to you too! I trust you slept well. Was I in your dreams?
It's strange to see such a pattern of deliberately visiting threads so that you can just dish out some childish hate
No, I merely responded to your own [slashdot.org]. Your shtick always makes life so easy...
Oh, and you should know the rules by now. Only one question per post.
Re: (Score:2)
No, I merely responded
Without a lick of context or on-topic substance of any sort. As usual. Because you're afraid of actually speaking on topic.
Re: (Score:1)
You are topic free... Just blabbing whatever FOX tells you to piss and moan about this week.
Re: (Score:2)
Re: (Score:1)
There you go again [deviantart.net]
Re: (Score:2)
Your post is a perfect example of what I'm talking about. You managed to get all the names in there, PC, SJW, special snowflake, professional victim...
Re: (Score:3)
Re: (Score:2)
Name calling is not an argument.
You are still just proving my point. Sorry, but you have no right not to be offended or made to feel bad about the things you say. If you say something, it's legitimate to criticize it. Simply claiming that the criticism is "PC" is merely an excuse, it does't actually present an argument supporting the statement.
I gave some specific examples. You don't challenge them. That suggests that you actually think some points you label as PC are legitimate.
first define "abuse" (Score:2)
without defining the vague indefinite meaning of word "abuse", this is bs.
and since this is media that deals with words, worst that can happen( if there is no other contact external to this media) is hurt feelings.
are such hurts, 'abuse'? (or even definable) given the highly subjective nature of emotions.
so this is bs.
this so called ai merely checks to see comments on yahoo news conform to rules of western 'liberal' elites. it is form of news censorship.
--
"a man who lies to himself is often the first to t
And so it has begun... (Score:4, Insightful)
ModBot: I am sorry, but your message was removed because of Violation of rule #157792 - negative opinion on political minority group.
ModBot: I am sorry, but your message was removed because of Violation of rule #151734 - negative opinion of product. Be fruitful!
ModBot: I am sorry, but your message was removed because of Violation of rule #191727 - hate speech: you voiced an opinion on criminals. Let's leave that to our law enforcement, right? Be well!
ModBot: I am sorry, but your message was removed because of Violation of rule #1 - Personal Opinion. We encourage our citizens to support each other, personal opinions are best kept to yourself, be well citizen!
Re: (Score:2)
So literally the only difference is that it's a bot doing the moderation. Posting threats on XBox Live is not allowed either way, read the TOS. Free speech unaffected.
Face it, you can't expect mainstream commercial platforms to give you a soapbox for absolutely anything you care to say. There are less popular sites like Voat and 4chan anf even Slashdot, or start your own.
Re: (Score:2)
Re: (Score:2)
There is room for both. It's hard to have a good discussion on an unmoderated forum, just look at the quality of debate on 4chan. That's why most of the popular sites are moderated, and people prefer them. It's why IRL debates have moderators.
Demanding every site accepts your extreme arguments is actually just demanding people who don't want to get into yet another argument about how the holocaust is fake be silenced, or at least drowned out. Just because someone doesn't want to deal with that doesn't make
Re: (Score:2)
You have been fined 1 credit for violation of the verbal morality code.
Who could have known that a Sylvester Stallone action movie would be so fucking prescient? It even predicted tablet computers and car AI's.
One More Nail In Yahoo's Coffin (Score:1)
Slashbros, freak the fuck out! (Score:1, Troll)
Yahoo, a private company, doesn't want people posting comments on its properties that it considers abusive. Somehow, that is anti-free speech and contrary to the God given First Amendment freedom to be a dick online. Slashbros demand their right to insult, harass, and threaten anyone they deem to be "PC" or a "SJW" in any forum, and will cry vainly at any attempt of a free market player to deny them that right.
This isn't the government mandating that Yahoo take these steps, mind you, and nobody's freedom of
Re: (Score:1)
Almost tautological? (Score:2)
If the new AI didn't outperform the old AI, it seems unlikely that they would have deployed it.
Let's focus on something else (Score:2)
I mean, Yahoo using an abuse-detection isn't news, since they had an older one in place that the new system is beating.
So then the news is that they set up a better machine learning algorithm with better training data, and the results were better? Color me shocked.
In other words, a tumblrina-complete AI. (Score:2)
Expect a whole lot of conservative speech to be in there or at least operating with Yahoo's INGSOCJUS bot.
Internet will have a field day with this AI. (Score:2)
They will teach it that the letter "a" by itself is an racist remark or something, and this will simply shut down the whole comment section being protected by the AI.