Hacker News Reader

Comment by joatmon-snoo | Hacker News Reader

People keep saying that Facebook should be doing more to block fake news, but here's the part that I don't get: why do you want Facebook to exercise even _more_ discretion over what kind of news people see?

That's to say nothing of the sheer difficulty of classifying something as "fake news". What's the threshold? How do you determine if something is fake news? Should stuff like Clickhole and The Onion - y'know, satire news - be considered fake news? What about sources like the Huffington Post, which masquerades as a newspaper but is literally a glorified blogger platform?

View on HN

Fighting fake news would not increase the degree to which Facebook shapes what people see. It would just adjust the criteria. Facebook already exerts total control over what its users see.

"Discretion" is not the right word, though, because Facebook is algorithmic. The problem is simply that the algorithm is optimizing for the wrong thing, which is causing harm.

It's like if you decided what to eat and drink solely by how sweet things taste. Each individual bite would be pleasurable. But the long-term health effects of eating nothing but sweets would be terrible.

Blocking is not the only answer BTW. Simple labeling would go a long way. The Onion would still be funny with a little "satire" tag.

It would just adjust the criteria from things Facebook thinks they'd want to see to things that powerful people think would tilt the users' political views in the right direction. That's a major change, and one that - by design - transfers a huge amount of power and control from individual Facebook users to the centralized approvers of permitted content.

Thinking about politics first is a backward way to look at the world.

We need to know the truth about facts. Only then can we make useful political decisions about those facts.

"The truth about facts" is extremely ambiguous. However, I'm sure the new Minitrue department at Facebook will solve this problem.

How does Facebook determine what facts are without politics?

Facebook fact checking materials before displaying them is fundamentally political, if only because the decision of filtering strategy is a political choice.

A news story that the Pope endorsed Trump is fiction and should be labeled as such.

That has political implications because everything does. But evaluating content based on those implications, rather than on the actual facts, is a categorical error.

We label fiction and nonfiction in libraries, book stores, movies, magazines, etc--basically every other medium for accessing content. Why not Facebook too?

How do you define fiction?

"The Pope endorses Trump" would be labeled as fiction. But what about "The Pope endorses [some of] Trump's views"? Or "The Pope is a bigger fan of Trump than you think"?

How about start with the most blatant examples... like spam detection

Facebook is not even addressing the most blatant, obvious fake news.

In fact, they do the opposite: they promote it to millions of people. Or more accurately, they permit their software to do so. And they claim they have no responsibility, even though they built the entire thing.

Facebook does not want to lose revenue from advertisers that submit fake news articles.

I saw the BS News extension it checks each link for citing known fake news sites in their article.

It is not practical as I keep seeing new fake news sites popping up from nowhere. Just register a domain, install Wordpress, write some articles and then pay for advertising on Facebook.

Fake news sites got Trump elected by saying false things like the Pope supports him. It is serious when people only use Facebook as their only news site and read and believe without checking the facts first.

If I were to make a fake news site detector I'd have it search Snopes and Politifact and others to see if it has been already reported as false.

Basically the way it ends up is a new way to censor the news using AI.

Suppose instead of a black list of sites they use a white list of trusted sites? Then block any new sites until they can verify them as trusted?

You didnt answer my question at all: how does Facebook, prior to the political decision of how to determine what facts are, know that the Pope endorsing Trump is fiction?

So... When have you stopped beating your wife?

Determining what facts are is not a political decision.

Decisions with political implications are not the same thing as political decisions. Huge categorical difference, in fact.

  How does Facebook determine what facts
  are without politics?

Perhaps they could determine what facts are by seeing if they're true or not?

For example, if an article states that Colorado pot shops accept food stamps [1] or that Dearborn in Michigan follows Sharia law [2] perhaps Facebook could determine those articles are satirical because they aren't true?

[1] http://www.denverpost.com/2014/01/14/fake-report-about-color... [2] http://nationalreport.net/city-michigan-first-fully-implemen...

ha, it's quite obvious if you check the front page http://nationalreport.net/

They may need to make things more obvious.

How does Facebook determine if something is true prior to the political choice of the method they'll use to determine truth?

You just moved words around, but my original question is still there: how do you without any political decisions determine what is true?

There is no single method for determining truth, thus which for a collective to use is fundamentally a political choice.

You and a sibling comment seem to be forgetting that we need to decide how we're going to determine truth before we go about doing it. In fact, you assume Facebook knows a fact (that potshops dont accept foodstamps), but how does Facebook know that is a fact or not?

> How does Facebook determine what facts are without politics

Facts are facts. The pope did not endorse Donald Trump. That is a fact.

How does Facebook, without making a choice about what method to use, determine that the pope didn't endorse Donald Trump?

If Facebook makes a choice of method, it's fundamentally political, as all group choices are.

Thus, Facebook can't escape the political nature of that task: epistemology requires a philosophical position, and a philosophocal position being adopted is a choice (for an organization), and all group decisions are inherently political.

That applies to everyone in life and you're already doing that by looking at an algorithmically curated feed. I'd much rather it at least tell me what's fake news so I don't waste my time.

Facebook itself is already a bunch of powerful people - the people who make the decisions about how content is served to users.It's already a centralized source of power.

Facebook already exercises this power by deciding how to determine what users "want" to see, i.e., what Facebook thinks users want to see.

I'm not sure how you could claim that individual Facebook users have much power at all. What they see is totally determined by the platform.

Facebook doesn't show people things Facebook thinks they'd want to see. Facebook shows people things that Facebook thinks will get them to spend more time on Facebook. There's a very important distinction between the two.

> The problem is simply that the algorithm is optimizing for the wrong thing, which is causing harm.

It's optimizing for the right thing, if you are Facebook: it's giving you more and more content you feel comfortable with, so you spend more time on the site and they can sell more ad space.

"Facebook already exerts total control over what its users see."

That bothers me a little. I have a page and Facebook somehow decides who and when a post is shown to a user who liked the page. It seems that Facebook shows it to a small group of people over an hours time or so and then depending on whatever metrics and interactions are made to a wider group or not.

I don't know if it is good or bad because some posts that do very well on Reddit and Hacker News do very poorly on Facebook while posts that do very well on Facebook can do very bad on Reddit and Hacker News.

Ultimately it's a very hard problem to solve.

Fake news, and its far more dangerous big brother, propaganda, can have an enormous effect on a society. Providing people with tools to identify real information from fake information seems like a great thing for those who want to use them.

But a lot of folks (whether they realize it or not) want to see fake news because it confirms their personal worldview (or: how everyone loves confirmation bias) and makes them feel good about being part of their in-group.

I suspect folks who push for censorship think that it will somehow ensure these types of folks don't consume fake news or propaganda (probably without realizing that they, too, have consumed fake news from time to time).

But the reality is fake news has existed for a very very long time. Hell, you could tune into MSNBC or Fox News and find stories that border pretty closely on fake news without even trying. And, again, they do it because it increases ratings. And it increases ratings because it makes people in their respective demographics feel good.

So what you're ultimately trying to fight is the human need for a sense of belonging, and the desire to feel good about their personal beliefs. Good luck with that...

Fake news is propaganda, full stop. While it is often more concerning when done by a state actor, there is no fundamental difference.

Yes. The problem is most of these companies that want to fight "fake news" would probably almost completely focus on "new sites" that have fake news, and perhaps a couple of other "more obvious" big sites like Breitbart.

But what happens when the fake news comes from CNN? Will Facebook block CNN or WSJ articles that spout lies and are blatantly incorrect?

The problem with this is that these days, you often also get the "real news" from smaller sites, like DemocracyNow, or others even smaller, that report on NoDAPL protests when none of the big news sites do. What if Facebook gets used to censoring these sites, though, even though they report the real news out there? And would we even know? After all, they are small sites, that few people would know about in the beginning.

Fake news is propaganda, full stop

The Onion might disagree with that. The line between one man's satire and another man's propaganda is a pretty thin one.

Not so thin at all. I think there is a pretty clean delineation between satirical sites like the onion (i.e. Social commentary) and the sort of "fake news" we are talking about here. These fake news posts are presented as factual, with the purpose of supporting a political or social goal - in other words propaganda.

Welcome to the original "editorial" question. Read up on "Yellow Journalism" for a history.

The answer to your rhetorical is "some people start from a position of reporting responsibility some start from a position of personal responsibility."

Lets presume there is a conversation between Alice and Bob on what is happening. Eve can hear both conversations and to some extent can ascertain when either Alice or Bob are communicating a falsehood or a misstatement to the other.

In the personal responsibility policy camp Eve doesn't do anything, it is up to Alice and Bob to evaluate the integrity and accuracy of the other's facts. In the reporting responsibility policy camp, if either Alice or Bob is communicating to a group (which is to say the communication is one to many) then Eve must inform the many when Alice states a falsehood or lies, conversely if Bob is the one who is communicating to the many then his falsehoods and/or misleading statements must be called out.

Facebook lands in the middle because both Alice and Bob are in fact communicating to lots of people (everyone on the feed, everyone reading the comments section what ever). As the "market maker" for this meeting of speakers and listeners, either Facebook is 100% responsible for the quality of the facts or 100% not responsible for the quality of the facts.

If you think of the information being exchanged as "product" and people publishing information as "sellers" and people reading that information "buyers" (even though there may be no financial transaction involved), there is quite a bit of experiential guidance around what the market owner must do versus the sellers and buyers in the market. The challenge is that people who might be suspicious being offered a Rolex watch at a flea market for a fraction of the price, seem more willing to believe a "news story" from an unaccredited news source which differs markedly from what the mainstream press is reporting. If they saw such news stories as 'cheap rolexes' it might activate the same sorts of suspicions they would get in a market when the price is too good to be true.

That's why you don't block; you flag. Let people make their own choices, but informed. Allow users to flag a link or post with tags such as Satire, Misleading, Controversial, Editorialized, Biased, Debunked, etc. Everyone else can still view the links and posts, but with some small indication of its trustworthiness. Though I feel like some tags, such as Debunked, should perhaps require a supporting link attached to a flag.

What about 'Deplorables'? Can we tag news stories as that?

Tagging by users can easily be gamed but it would be a useful signal for an algorithm.

> Allow users to flag a link or post with tags such as Satire, Misleading, Controversial, Editorialized, Biased, Debunked, etc.

The entire problem is that most people don't know or realize when articles qualify for any of these criteria.

Are the ads also getting a "biased"-tag? It also stifles media competency. I think a better approach would be to give a daily or weekly summary of questionable Articles that where opened by the user.

It's easy to be outraged rather than actually thinking about the problem.

HEADLINE:"FACEBOOK NEWS FEED POPULATED BY BIASED EDITORS"

INTERNET:"MACHINE LEARNING IS UNBIASED WHY ISN"T FB USING ML?" -- HEADLINE:"FACEBOOK NEWS FEED POPULATED BY FAKE NEWS BECAUSE ML ISN'T DIFFERENTIATING NEWS TYPES"

INTERNET:"FACEBOOK NEEDS TO CONTROL WHAT NEWS GETS ON THE NEWS FEED" --- HEADLINE:"FACEBOOK CENSORING CONTROVERSIAL NEWS" INTERNET:"FACEBOOK NEEDS TO HIRE EDITORS THAT WON'T CENSOR THINGS"

GOTO 1

I don't think Facebook needs to control what gets on the news feed.

They could have avoided all of that by not interfering.

All they need to do is nothing. Don't favor some news, don't block fact-checkers. Just let me see everything my friends post.

Great point; the question then becomes why is this interference necessary? Largely due to ads. Instagram made their feed algorithmic recently for this reason.

The gray zone expands even more with the lesser-known satire sites, which sometimes have a mix of articles that are obviously satire, and some that aren't.

For example, I've seen a few articles shared as truth from "World News Daily Report," which played it so straight that I would have considered it "fake news" if I hadn't seen the satire disclaimer tucked at the bottom of the page. Other articles on that same site are much more obviously satirical though. I tend to blame this on poor writing rather than malice, though you can probably make a decent counterargument.

Some sort of "possibly false" notice attached to links like these, rather than just a "block or not" decision, seems like it would strike a nice balance here.

Also, the people making loud, direct calls for censorship are mostly journalists, writers, and opinion-makers. Seems like great fodder for views and clicks. Perhaps there's much for them to gain by whipping up drama (and potentially alienating competition).

No one thinks it would be a simple task, but if you consider it a threat to democracy you can understand why many want a solution.

Blatant lies intended to scare people into voting a certain way are a problem. How much of a difference is there really between "vote X or I'll kill your family" and "vote X or someone else will kill your family" if the intent is the same?

> How much of a difference is there really ...

... a pretty damn big difference there. Not the best analogy you could have used.

Facebook already exercises a lot of discretion -- they reorder posts on your timeline to drive engagement, they promote related links, they have the Trending bar, etc. People aren't asking them to do more. They're asking them to do what they're doing now more responsibly.

I couldn't agree more. I don't think Facebook should necessarily block fake news (however you want to define it) but it might be nice to let me know before I share a story that they think it's satirical because of reason A and reason B or that it's inaccurate for other reasons.

This is good for me because I don't look dumb for sharing something that's incorrect and it's good for Facebook because now they know even more about me (that I love satire but hate fake news, for example).

>>"People keep saying that Facebook should be doing more to block fake news but ... why do you want Facebook to"

For the same reasons why I want my email provider to block spam or for my search engineer provider to block shady viagra ads.

> why do you want Facebook to exercise even _more_ discretion over what kind of news people see?

The argument is that, whether we like it or not, Facebook has become the hub for millions of people and these people share articles with each other. Media has become decentralized, and although the libertarian approach of do-nothing sounds nice, we just found out that in practice it doesn't work very well. You thought the Reddit fiasco with the marathon bomber was bad? The equivalent is happening a thousand times over with fake news spreading faster than the news that debunks it.

What's your metric for communication platform full of user generated content 'working well'?

I'm not trying to be snarky, I don't see how this can possibly be a failing of Facebook unless you really believe that an algorithm that ranks posts can actually fix the human problems that underlie this issue.

When I said that it wasn't "working well", I was referring to the epidemic of fake news. It turns out that, for better or worse, a significant amount of the population is using Facebook for its news source. We can say that they're silly and shouldn't be using Facebook for news, but they are and they're not going to stop until a systemic solution is put into place.

I think it is because people are generally NOT saying that they should necessarily BLOCK things, but instead label them appropriately.

THIS IS SATIRE: at the top of ONION articles. THIS IS FAKE: on top of us-political-truth-grabber-buzz.com

maybe OPINION gets added to HUFFPO

I think that would likely be enough to satisfy most of the needs.

... Although while we're talking blue sky / greenfield ... maybe we add something like:

   Here's How Ridiculously Expensive It Is To Replace A Windshield On An Aston Martin |[ $4414.41 ]

to all that shitty clickbait out there :)

YOU WON'T BELIEVE WHAT THIS MAN DID FOR HIS DOG! [He dressed up like it's favorite toy and it was adorable]

God that sounds amazing. I'd want to see the dog video, but I refuse to click on principle.

Can crowdsource summaries and make a chrome plugin?

The crowdsourcing is almost already there: https://www.facebook.com/StopClickBaitOfficial/

As long as it is public as to what they are blocking or flagging, then it's better than not doing it at all. The bigger problem is people may not either care or believe something is fake news, just because someone highlights it as such. It would be nice if there were questionnaires with people's voting, so post-election we could see how much understanding or knowledge different groups have; you couldn't, at least initially, put any weight on what someone's vote is worth based on this of course.

>but here's the part that I don't get: why do you want Facebook to exercise even _more_ discretion over what kind of news people see?

This one is easy: people in these parts think that the fake news is one-sided, and that it got a president they don't like elected. So now fake news is a major issue.

And the solution is, obviously, technology, and specifically machine learning because that's the solution to everything these days.

No. The solution is to setup a Ministry of Truth like what Facebook is trying to do. People are too stupid and must be protected from reading #FakeNews because they can not think for themselves. Let the corporations and the government think for them.

I suggest Facebook block all types of content… they can start by blocking all personally-identifiable information. Short of that, they should be responsible and shut down their irresponsible service.

Wouldn't it be reasonable to force facebook and similar social network sites to be transparent about how they rank articles? And in combination with this make reasonable legislation on content responsibility.

BuzzFeed would certainly love that. How to maximize your facebook score, complete with regress test framework.

Funny that they managed to block the fact-checker.

I'd be happy if they did nothing, and simply let me read what I want. But this is the worst situation.

It's easy to verify that a plugin modifies the Facebook site and block links to it. It's much harder to tell whether the list of fake news sites it uses is itself fake or the plugin is a security issue, let alone whether particular articles are truthful.

> It's easy to verify that a plugin modifies the Facebook site

No, it doesn't. That's a serious abuse of the language. What you describe would be hacking into Facebook.

What the plugin does is help me not load content I don't want.

The browser is my user-agent and I'm instructing it how to display the content delivered to me.

> It's much harder to tell whether the list of fake news sites it uses is itself fake

Exactly, and yet they're trying harder to block a fact checker than to block the bad news.

If they just backed off on both I could see the bad news and have my plugin flag it.

> or the plugin is a security issue,

Not even close. It's not like Facebook is cleaning up malware plugins.

They're only targeting this one because it threatens their editorial control of my feed.

which they are completely within their rights to do

Well, actually, perhaps not. That's the point. Past a certain size, if you're sleazy and manipulative we're going to control your actions.

But even if, how is that "sure, but you can't legally stop them" a defense? How is it one you feel compelled to make?

Especially because I can technologically stop them. They can block one plugin, and detect one set of blocks, but it's a game we've played many times before and the defense always loses.

Reply on news.ycombinator.com