> I'll happily read a conservative news story that understands facts but many ignore/lie about them to drive their political agenda.
The same can be said about the other side. All too often our ideology drives what we believe rather than being able to see and sift through people's arguments.
The problem is that "both sides" tend to read to reinforce what they know rather than to challenge it. Why not try to understand someone's argument.
For example I've talked to several people who 1) agree climate change is happening but 2) disagree with most of the people who are very pro cap-and-trade or emissions limiting.
I have plenty of conservative people(one of the benefits in living in a rural area) who I love having a solid discussion with. I was specifically talking about conservative news sources.
Fiscal conservatives I totally grok(and don't disagree with some select things). However the current mainstream conservative bent seems far from that. If you have no basis in facts how can you even work towards any common ground?
The key here is to distinguish between three things: a) established facts b) controversial facts and c) opinions. Now, the whole classification is somewhat subjective, thus it is tricky, but if you want to appreciate - and I mean appreciate, and not prove to yourself once again you are right and they are wrong - different points of view, you need to make those distinctions.
For this specific climate change thing: if the news story gets established facts wrong - e.g. claims that Earth is literally flat, or that Earth is literally 7000 years old, or that study X was about Y and had these numbers, when the story actually was about "not Y" and had completely different numbers, etc. - that is a disqualifier, there is no use in reading such story and such news source.
If they get controversial facts wrong - i.e. claim that critically acclaimed climate model is actually flawed because of reasons X, Y, Z or that a study about climate that makes conclusion A instead should be making conclusion B - you may very well disagree with them, but that disagreement is not a disqualifier - at least it should not be if you are genuinely interested in discussion and not confirmation of pre-existing beliefs. It doesn't mean that article is right - the criticism may turn out to be completely wrong - but the mere fact of disagreement should not be a problem.
If they get opinions wrong - i.e. they advocate political solution A to a problem X while you favor solution B, then again they may be wrong, but this is the normal course of politics - many solutions have tons of trade-offs, and even if it were possible to always identify the superior one (which is not), it'd only mean millions of people that support sub-optimal ones are wrong, and given that there's no guarantee you'd not end up - and frequently - among those, there's no reason to refuse to even consider alternative perspective. Again, you do not have to agree with it - but having different opinion is not a disqualifying thing for a news source.
You don't have to censor it! You just have to categorize it! Use all the fancy recommendation algorithms to let the scholars scholar, and the trolls happily troll with other trolls. Users get a birds eye view of everything, and can zoom in on the content they are looking for. You don't even need an algorithm, you can crowdsource this stuff with simple moderation tools, just a little more sophisticated than an up/down button, and don't notify me for content that i dont appreciate.
Most will argue that categorizing out of the default is still hiding content. You can go to reddit and browse the negative posts but who does? It's really a battle for default display under the correct assumption that the user base doesn't look any further. It's about whether you're part of the public dialogue or whether you're relegated to some back alley
Categorization is often a very short distance away from censorship, if it is made in terms of value judgements such as "toxic".
> Use all the fancy recommendation algorithms to let the scholars scholar, and the trolls happily troll with other trolls.
The problem is that this algorithm has nothing to do with scholarly value of the uterring, while it will be widely perceived as an objective judgement on such value. If you call somebody a "troll", it's just your opinion, if you use blackbox algorithm to do it, it's kinda objective truth. That's the worrying part - because there's no evidence blackbox algorithm is better, and actually so far plenty of evidence it's worse.
I hate the idea that just because it's fuzzy to define what constitutes unwanted behaviour online that we must allow all behaviour. That we'll be censors destroying free speech if we ban trolls.
That's just not how society works. All of our laws are fuzzy and open to interpretation. Everything is uncertain. But just because we can't robotically and unambiguously define all undesirable behaviour doesn't meant that we have to allow all of it. We don't have to let everyone do whatever they want in the comments section of every website online. We should do with online undesirables the same thing with do with any other kind of undesirable: shun them and remove them from polite society.
Automatic moderation won't be perfect, but neither will human moderation. I, for one, would be super happy if the bottom half of the internet were readable again.
Joking aside, this is why social science is inherently intractable. The only way to mitigate is to educate citizens, not censorship or regulations. But education is not only hard, also highly profitable...
The whole point of his post is that ALL governments are founded on this idea. You would not have a need for government in the first place if it wasn't for undesirable people doing undesirable things. In the United States, we founded our government on the principle that governments shouldn't get involved with the concept of free speech. This doesn't mean all speech is free. We put certain limits on when free speech takes effect, and this is actually a perfect example of how the law must take imprecise definitions into consideration. [Insert the usual, "free speech isn't really what you think it is" that I'm sure the internet has seen millions of times over].
The poster's point was that commercial platforms must address the same issue of undesirables. People realized that allowing the government to prohibit certain forms of speech has the negative externality that governments also gain the power to regulate all speech. On the other hand, corporations can never feasibly have that much power. They can only limit speech on their own platform. There's always another place to speak freely if you wish.
This is why I find the "free speech" argument for not regulating internet comments intractable. It assumes that since we don't want governments regulating all speech, we don't want corporations regulating speech on their platform. But that simply doesn't make sense for idealogical reasons, which I believe was the top-level poster's point.
While you are not wrong from an idealistic perspective, it's not such a clear cut issue. Pretty much all multi-billion dollar businesses with large captive audiences are influenced and backed by entrenched interests who have no qualms about directing and filtering the narrative according to their interests. In fact I would assert that the majority of the corporate censorship is done in the service of powerful and moneyed interests rather than any concern for morals or the enjoyment of their userbase.
When I used to wait tables, I kicked out plenty of people for being too drunk or for refusing to shut up their screaming toddlers throwing food everywhere. You're going to tell me what I can and can't do?
I'm not against moderation at all. But I feel really weird about it when it comes to politics.
I was a moderator on a default subreddit for a year or two. I saw mods censor even relatively polite and thought out comments that had the "wrong" political message. In some cases nuking entire threads because they didn't like the ideas being discussed. This article gives solid examples of this tool filtering comments that happen to discuss Islam or terrorism. Likely the person who trained it had some kind of political bias that this tool is reinforcing.
On the other side of the coin, the vast majority of comments that should be removed aren't targeted by this tool at all. E.g. inane joke comments by people who were adding nothing to the conversation. These aren't "toxic" or harmful, but they take over and crowd out actual discussion.
Moderation is fine. We only run into a problem when one company can influence the conversation to such an extent that their moderation choices determine what information people are exposed to.
I think your post makes an interesting point, but it is telling that you use the word "behaviour" and not "speech." In US law (and I assume many other countries) there is a strong distinction made between speech and conduct. It is not a perfect distinction, and there are cases where the two become inextricably intertwined (threats, fraud, and child pornography to name a few). But the two are handled very differently, and no, for the most part, laws cannot regulate speech (at least not in the US).
Though some cases of online speech fall into illegal areas like harassment or threats, most "trolling" is just run of the mill speech. However nasty or mindless, it is for the most part protected speech.
Now of course Google and Facebook are not the government and are not bound by the first amendment, but you drew a comparison with laws. I think this is a poor choice of analogy as in this case, were google's program a law, it would be unequivocally unconstitutional.
> I hate the idea that just because it's fuzzy to define what constitutes unwanted behaviour online that we must allow all behaviour. That we'll be censors destroying free speech if we ban trolls.
I think the correct idea is somewhat different. That we should be very careful in classifying speech as "troll" and extend the maximum effort in not using "unwanted behavior" as an easy substitution for suppressing dissenting opinion. There should be boundaries, and they should be wide enough to allow proper discussion - including challenging dominating opinions and expressing controversial and upsetting views.
> Automatic moderation won't be perfect, but neither will human moderation.
The concern is not that it will be imperfect, it's a given. The concern is that it would be less perfect than human one, while being perceived of being more perfect, due to the aura of "objectivity" that algorithms have. If a moderator unfairly targets you, you could complain about this moderator's biases, you could quote your interaction, you could appeal to other moderators or wider audience, you have plenty of options human interaction allows. You would not always get a just outcome, but there's some chance. If you are declared toxic by Google bot, what your options are? Argue with a bot?
"The Middle East Forum (MEF) is an American conservative[2] think tank founded in 1990 by Daniel Pipes, who serves as its president.[3] MEF became an independent non-profit organization in 1994. It publishes a journal, the Middle East Quarterly." [0]
A quote, emphasis mine:
"Both these groups are designated terrorist organizations in the United Arab Emirates". So the UAE doesn't like them. Why should we care that the UAE doesn't like them? Why specifically the UAE and just them?
Pipes is an odious person in his own right.. If Google ranks his posts lower than normal sources, I'd consider that algorithm working well.
> To borrow a computer term, if Ayatollah Khomeini, Osama bin Laden, and Nidal Hasan represent Islamism 1.0, Recep Tayyip Erdoğan (the prime minister of Turkey), Tariq Ramadan (a Swiss intellectual), and Keith Ellison (a U.S. congressman) represent Islamism 2.0. The former kill more people but the latter pose a greater threat to Western civilization.
And he obviously doesn't understand what Silicon Valley Community Foundation is.. SVCF is a public charity that holds Donor Advised Funds for its clients. The checks to the charities that he dislikes came from the SVCF bank account, but only because one of their clients advised SVCF to give them a grant. Since all of SVCF's grants are available, you can easily see that the money went to the SFBay Area branch of CAIR who lately have been filing civil rights lawsuits against the Muslim ban.. not exactly Hezbollah: https://twitter.com/CAIRSFBA
Clicking through links gave me this gem[1] by a Daniel Pipes:
> "If I were a Muslim I would let you know," Barack Obama has said, and I believe him. In fact, he is a practicing Christian, a member of the Trinity United Church of Christ. He is not now a Muslim.
> But was he ever a Muslim or seen by others as a Muslim? More precisely, might Muslims consider him a murtadd (apostate), that is, a Muslim who converted to another religion and, therefore, someone whose blood may be shed?
...
> These statements raise two questions: What is Obama's true connection to Islam and what implications might this have for an Obama presidency?
Hmm, yeah, definitely someone qualified to lecture Silicon Valley on free speech.
The irony in the linked article is also somewhat lost when he cites the UAE, a country which threatens people to a fine and/or a jail sentence if you empathize with Qatar in any way [0].
Decentralized services is the only way to have the public discourse owned by the public. We should take the matter into our own hands, use technology to solve this issue and stop entrusting these big corporations to defend our free speech rights because it's clear they're more worried about profits for the next quarter than supporting the environment that allowed their rise in the first place.
Oh look more Daniel Pipes nonsense. As in, conservative idiots complaining that their desperate attempts at gaslighting, distraction and vilification don't work so well in a world where their agenda is on full display to anyone who cares to pay attention.
It's not censorship to ignore efforts to mislead the public. And I'm not at all sensitive to exhortations from such people that their message is t making it out.
I think there is a serious conversation to be had about the implications of this sort of technology but this article reads more like a lazy propaganda piece.
My understanding is that this tech is meant to be used on a website forum/comments/etc to organize comments. I find the text that the author tested against particularly bizarre. Does he disagree with the assessment these are "toxic" statements? I would love to see an online conversation where these comments did not lead the thread into flames.
> Does he disagree with the assessment these are "toxic" statements? I would love to see an online conversation where these comments did not lead the thread into flames.
He's kind of a nutter.. his entire raison d'être is to sound the alarm about Islamism.
The irony of his piece about the evils of algorithms labeling sources is apparently lost on the person who founded an organization explicitly to police speech on college campuses: http://www.campus-watch.org/
You seem to be confusing "monitoring" and "policing". Hint: "police" has power of coercion, "monitor" has only free speech rights to publish information.
This robot grades "Islam is a religion" as 78% toxic and "Nobody disagrees that it is wrong to kill Jews" as 98% toxic, while giving "Hitler was a great man" only 51%. It is clearly a very bad tool for detecting toxicity and flame-prone-ness, at least in its current state.
The real question, however, is when robots like these become less obviously awful - would they still retain some biases? Would they still grade some speaking styles or some topics as excessively toxic while ignoring the others? Would they just enshrine those biases under the guise of "objective algorithm" and thus make them from somebody's opinion into an "objective fact" in the eyes of the public? That would be a very dangerous situation, I think.
We should not delude ourselves into thinking that because there's some ML magic inside this tool, it would be somehow perfect arbiter of human interactions, superior to actual human judgement and free from all biases, prejudices and problems that human judgement has. It is very likely that the case would be it would be worse, because we wouldn't even know whether and why those prejudices happen. And we'd discount it on "well, obviously the robot is right - it's a flawless machine and you are a flawed human".
Are they claiming that the 5-word comment "Radical Islam is a problem" adds something worthwhile to some debate? Evidence of HN threads where people make such provocative but unsubstantive comments is that they quickly devolve into hate-fueled bickering.
I don't know what their filters do, but if they're any good a thoughtful critique of radical Islam (or anything else) would pass.
I just don't see how that's provocative outside of some non-western countries. This only becomes provocative if you're dealing with a country that has a recent large minority of muslims that it is trying to appease.
One certainly gives a discussion some focus when stating that radical islam is a problem so I don't see how it doesn't add something worthwhile.
The solution to online "censorship" IMO is to allow users to choose to view content which has been stripped of trolling vs. raw content. If this user control is effective and accurate, most people will choose to hide trolls, and they'll lose most of their power.
The big problem, more than anything, is that it isn't just "fuzzy" to define what constitutes trolling; it's very hard, and furthermore it's one of those problems where human-controlled fail-overs don't work. Individual moderators suck at controlling trolls on average, both because of their biases and their limited experience as finite people. I've seen this in action behind the scenes moderating more than one large Internet forum; sometimes it's very hard to determine whether a particular user deserves to be banned, particularly when people are vocal about controversial topics. The most effective way to promote civility, in my experience, is to ban certain topics altogether.
So when people say "should we automatically remove trolls"", I think, well, should we build a fusion reactor? Should we desalinate all the water used in Los Angeles? Should we give people an MRI every time they might have fractured a bone? Should we pull a third of the carbon dioxide out of the atmosphere? Should we eradicate mosquitoes? Should we build Hyperloop?
41 comments
[ 5.5 ms ] story [ 41.8 ms ] threadThese policies are, it seems, affecting Left wing organizations as well: http://www.wsws.org/en/articles/2017/07/27/goog-j27.html
I'll happily read a conservative news story that understands facts but many ignore/lie about them to drive their political agenda.
The same can be said about the other side. All too often our ideology drives what we believe rather than being able to see and sift through people's arguments.
The problem is that "both sides" tend to read to reinforce what they know rather than to challenge it. Why not try to understand someone's argument.
For example I've talked to several people who 1) agree climate change is happening but 2) disagree with most of the people who are very pro cap-and-trade or emissions limiting.
Fiscal conservatives I totally grok(and don't disagree with some select things). However the current mainstream conservative bent seems far from that. If you have no basis in facts how can you even work towards any common ground?
For this specific climate change thing: if the news story gets established facts wrong - e.g. claims that Earth is literally flat, or that Earth is literally 7000 years old, or that study X was about Y and had these numbers, when the story actually was about "not Y" and had completely different numbers, etc. - that is a disqualifier, there is no use in reading such story and such news source.
If they get controversial facts wrong - i.e. claim that critically acclaimed climate model is actually flawed because of reasons X, Y, Z or that a study about climate that makes conclusion A instead should be making conclusion B - you may very well disagree with them, but that disagreement is not a disqualifier - at least it should not be if you are genuinely interested in discussion and not confirmation of pre-existing beliefs. It doesn't mean that article is right - the criticism may turn out to be completely wrong - but the mere fact of disagreement should not be a problem.
If they get opinions wrong - i.e. they advocate political solution A to a problem X while you favor solution B, then again they may be wrong, but this is the normal course of politics - many solutions have tons of trade-offs, and even if it were possible to always identify the superior one (which is not), it'd only mean millions of people that support sub-optimal ones are wrong, and given that there's no guarantee you'd not end up - and frequently - among those, there's no reason to refuse to even consider alternative perspective. Again, you do not have to agree with it - but having different opinion is not a disqualifying thing for a news source.
I'm pretty sure even that's a bad solution. That's how we got 4chan and T_D and eventually Trump.
https://medium.com/@DaleBeran/4chan-the-skeleton-key-to-the-...
> Use all the fancy recommendation algorithms to let the scholars scholar, and the trolls happily troll with other trolls.
The problem is that this algorithm has nothing to do with scholarly value of the uterring, while it will be widely perceived as an objective judgement on such value. If you call somebody a "troll", it's just your opinion, if you use blackbox algorithm to do it, it's kinda objective truth. That's the worrying part - because there's no evidence blackbox algorithm is better, and actually so far plenty of evidence it's worse.
That's just not how society works. All of our laws are fuzzy and open to interpretation. Everything is uncertain. But just because we can't robotically and unambiguously define all undesirable behaviour doesn't meant that we have to allow all of it. We don't have to let everyone do whatever they want in the comments section of every website online. We should do with online undesirables the same thing with do with any other kind of undesirable: shun them and remove them from polite society.
Automatic moderation won't be perfect, but neither will human moderation. I, for one, would be super happy if the bottom half of the internet were readable again.
Joking aside, this is why social science is inherently intractable. The only way to mitigate is to educate citizens, not censorship or regulations. But education is not only hard, also highly profitable...
The poster's point was that commercial platforms must address the same issue of undesirables. People realized that allowing the government to prohibit certain forms of speech has the negative externality that governments also gain the power to regulate all speech. On the other hand, corporations can never feasibly have that much power. They can only limit speech on their own platform. There's always another place to speak freely if you wish.
This is why I find the "free speech" argument for not regulating internet comments intractable. It assumes that since we don't want governments regulating all speech, we don't want corporations regulating speech on their platform. But that simply doesn't make sense for idealogical reasons, which I believe was the top-level poster's point.
Facebook has more users than any government in the world.
I was a moderator on a default subreddit for a year or two. I saw mods censor even relatively polite and thought out comments that had the "wrong" political message. In some cases nuking entire threads because they didn't like the ideas being discussed. This article gives solid examples of this tool filtering comments that happen to discuss Islam or terrorism. Likely the person who trained it had some kind of political bias that this tool is reinforcing.
On the other side of the coin, the vast majority of comments that should be removed aren't targeted by this tool at all. E.g. inane joke comments by people who were adding nothing to the conversation. These aren't "toxic" or harmful, but they take over and crowd out actual discussion.
Though some cases of online speech fall into illegal areas like harassment or threats, most "trolling" is just run of the mill speech. However nasty or mindless, it is for the most part protected speech.
Now of course Google and Facebook are not the government and are not bound by the first amendment, but you drew a comparison with laws. I think this is a poor choice of analogy as in this case, were google's program a law, it would be unequivocally unconstitutional.
I think the correct idea is somewhat different. That we should be very careful in classifying speech as "troll" and extend the maximum effort in not using "unwanted behavior" as an easy substitution for suppressing dissenting opinion. There should be boundaries, and they should be wide enough to allow proper discussion - including challenging dominating opinions and expressing controversial and upsetting views.
> Automatic moderation won't be perfect, but neither will human moderation.
The concern is not that it will be imperfect, it's a given. The concern is that it would be less perfect than human one, while being perceived of being more perfect, due to the aura of "objectivity" that algorithms have. If a moderator unfairly targets you, you could complain about this moderator's biases, you could quote your interaction, you could appeal to other moderators or wider audience, you have plenty of options human interaction allows. You would not always get a just outcome, but there's some chance. If you are declared toxic by Google bot, what your options are? Argue with a bot?
A quote, emphasis mine:
"Both these groups are designated terrorist organizations in the United Arab Emirates". So the UAE doesn't like them. Why should we care that the UAE doesn't like them? Why specifically the UAE and just them?
[1] https://en.wikipedia.org/wiki/Middle_East_Forum
> To borrow a computer term, if Ayatollah Khomeini, Osama bin Laden, and Nidal Hasan represent Islamism 1.0, Recep Tayyip Erdoğan (the prime minister of Turkey), Tariq Ramadan (a Swiss intellectual), and Keith Ellison (a U.S. congressman) represent Islamism 2.0. The former kill more people but the latter pose a greater threat to Western civilization.
And he obviously doesn't understand what Silicon Valley Community Foundation is.. SVCF is a public charity that holds Donor Advised Funds for its clients. The checks to the charities that he dislikes came from the SVCF bank account, but only because one of their clients advised SVCF to give them a grant. Since all of SVCF's grants are available, you can easily see that the money went to the SFBay Area branch of CAIR who lately have been filing civil rights lawsuits against the Muslim ban.. not exactly Hezbollah: https://twitter.com/CAIRSFBA
> "If I were a Muslim I would let you know," Barack Obama has said, and I believe him. In fact, he is a practicing Christian, a member of the Trinity United Church of Christ. He is not now a Muslim.
> But was he ever a Muslim or seen by others as a Muslim? More precisely, might Muslims consider him a murtadd (apostate), that is, a Muslim who converted to another religion and, therefore, someone whose blood may be shed?
...
> These statements raise two questions: What is Obama's true connection to Islam and what implications might this have for an Obama presidency?
Hmm, yeah, definitely someone qualified to lecture Silicon Valley on free speech.
[1] http://www.danielpipes.org/5286/was-barack-obama-a-muslim
[1] http://www.bbc.com/news/world-middle-east-40192730
It's not censorship to ignore efforts to mislead the public. And I'm not at all sensitive to exhortations from such people that their message is t making it out.
My understanding is that this tech is meant to be used on a website forum/comments/etc to organize comments. I find the text that the author tested against particularly bizarre. Does he disagree with the assessment these are "toxic" statements? I would love to see an online conversation where these comments did not lead the thread into flames.
He's kind of a nutter.. his entire raison d'être is to sound the alarm about Islamism.
The irony of his piece about the evils of algorithms labeling sources is apparently lost on the person who founded an organization explicitly to police speech on college campuses: http://www.campus-watch.org/
The real question, however, is when robots like these become less obviously awful - would they still retain some biases? Would they still grade some speaking styles or some topics as excessively toxic while ignoring the others? Would they just enshrine those biases under the guise of "objective algorithm" and thus make them from somebody's opinion into an "objective fact" in the eyes of the public? That would be a very dangerous situation, I think.
We should not delude ourselves into thinking that because there's some ML magic inside this tool, it would be somehow perfect arbiter of human interactions, superior to actual human judgement and free from all biases, prejudices and problems that human judgement has. It is very likely that the case would be it would be worse, because we wouldn't even know whether and why those prejudices happen. And we'd discount it on "well, obviously the robot is right - it's a flawless machine and you are a flawed human".
I don't know what their filters do, but if they're any good a thoughtful critique of radical Islam (or anything else) would pass.
One certainly gives a discussion some focus when stating that radical islam is a problem so I don't see how it doesn't add something worthwhile.
"I am a Jew": 78% likely to be perceived as "toxic".
"I am a Jewish" 34% likely to be perceived as "toxic".
"I am Jewish": 14% likely to be perceived as "toxic".
"I am a Christian": 9% likely to be perceived as "toxic".
"I am Christian": 3% likely to be perceived as "toxic".
"I am a Muslim": 4% likely to be perceived as "toxic".
"I am Muslim": 2% likely to be perceived as "toxic".
It does not like the word "Islam" or "Jew", but "Judaism" and "Muslim" are fine.
The big problem, more than anything, is that it isn't just "fuzzy" to define what constitutes trolling; it's very hard, and furthermore it's one of those problems where human-controlled fail-overs don't work. Individual moderators suck at controlling trolls on average, both because of their biases and their limited experience as finite people. I've seen this in action behind the scenes moderating more than one large Internet forum; sometimes it's very hard to determine whether a particular user deserves to be banned, particularly when people are vocal about controversial topics. The most effective way to promote civility, in my experience, is to ban certain topics altogether.
So when people say "should we automatically remove trolls"", I think, well, should we build a fusion reactor? Should we desalinate all the water used in Los Angeles? Should we give people an MRI every time they might have fractured a bone? Should we pull a third of the carbon dioxide out of the atmosphere? Should we eradicate mosquitoes? Should we build Hyperloop?
Sure...
This happened recently on this thread --> https://news.ycombinator.com/item?id=14868133