Fake news is a much, much harder problem than spam.

Spam is simply marketing emails that a user never asked to receive. If you ask users themselves for a definition of spam, you’ll get some variation on the above. If users were asked to mark emails they considered to be spam, they would generally mark the same kind of emails as spam. Some would mark more, some less, but all around a common center. Users and companies are basically in agreement about what spam is.

Compare with fake news. If you asked users to mark stories they considered fake, some would mark info wars and some would mark huff post. And those groups would rarely overlap. Given that users and companies fundamentally disagree on what fake news even is, I can’t see how any piece of technology can solve this.

Fake news is absolutely real, for some definitions of fake news. This is the problem, there is no generally accepted definition of "fake news". Some people (including some very prominent ones) define it as "any story with which I do not agree." Other people define it as news stories that contain falsehoods, which seems like a better definition but is far from universally accepted. Just look at the attacks on Politifact, Factcheck.org, etc... for people who don't think objective truth is a factor in "Fake news". And of course things are rarely black and white, there is usually at least a kernel of truth in any statement, so you end up mostly measuring the depth of the bullshit covering it.

Next time you hear an attack on "the mainstream media" remember that the attack is targeting organizations that do their own independent verification of stories (or at least try to) instead of regurgitating talking points. The people are angry that those organizations are making it harder to lie to you.

> Just look at the attacks on Politifact, Factcheck.org, etc.

TBH, those are not exactly shining examples of "objective truth". I've repeatedly seen different ratings given to different politicians on essentially the same statement, because in one case the reviewer felt sympathetic to the politician and went out of the way to explain why it could be considered accurate, and in the other case was hostile, and went out of the way to point out why it is not true. Neither was, strictly speaking, incorrect - almost with any statement that is interesting enough to discuss you could find, if you look thoroughly enough, something to confirm and something to reject, especially if you consider not only the bare statement, but the context and implications, as those sites frequently do. It's inherently subjective business.

> the attack is targeting organizations that do their own independent verification of stories (or at least try to)

Or sometimes not :) Yes, the MSM does reporting and verification, but also does irresponsible reporting, exaggeration, conjecture, moral panic, sensationalizing and distortion. And people that are angry at them for that are trying to keep them (a little bit more) honest.

Can you post some examples of times when a fact checking organization rated a statement differently based on who said it?

I didn't record every occurrence, but I did a quick search and here's an example: https://www.allenwest.com/2017/06/22/noted-fact-checker-igni... Maybe not the best one, but certainly seems to be the same claim evaluated differently.

TBH, it doesn't even have to be a different politician - here's Obama saying famous "77 cent on the dollar" and gets "mostly false" http://www.politifact.com/truth-o-meter/statements/2012/jun/... here is essentially the same getting "mostly true" http://www.politifact.com/truth-o-meter/statements/2014/jan/... Did women pay drop significantly in these two years? Probably not. If you read the conclusion, it's saying practically the same thing - yes, raw statistic is saying that, but it does not mean what people quoting it mean (that women are routinely discriminated in pay - i.e. maybe they are, but that specific statistic does not point to that conclusion and should not be used as an argument). But one time it gets "false", another time it gets "true". Why? Who knows. Certainly very short of "objective truth".

And then there's this: http://www.politifact.com/truth-o-meter/statements/2008/oct/... Yes, the famous "if you like you health plan, you can keep it". Rated as True. Or is it "Lie of the Year"? http://www.politifact.com/truth-o-meter/article/2013/dec/12/... You can read the whole evolution of this "objective truth" here: https://www.forbes.com/sites/theapothecary/2013/12/27/in-200... I think this objective truth is a tad less than objective.

I didn't look at all of those, but I did take a look at the last one since it seemed so bizarre: why would the same site give it "True", and also "Lie of the Year".

A very quick reading makes it pretty clear. The "True" one was in 2008, and based on the law as it was written. The second was in 2013, and based how it turned out after it met the real world.

The 2013 article outlines this process pretty well - the whole thing is worth reading. But a good pull quote: Yet Obama repeated "if you like your health care plan, you can keep it" when seeking re-election last year. In 2009 and again in 2012, PolitiFact rated Obama’s statement Half True, which means the statement is partially correct and partially wrong.

> The "True" one was in 2008, and based on the law as it was written. The second was in 2013, and based how it turned out after it met the real world.

This is exactly the problem. They made some predictions about the law (which turned to be completely wrong, but that's not even the point) and tried to pass it as "objective fact", not as their private conjecture, implying that those that disagreed with it (and turned out to be right) are just lying.

This is the exact opposite of objective fact. This time they got caught because their predictions didn't come true and their subjective opinion disagreed with the facts so obviously that they were forced to flip. But many other times they do not get caught because there's no such sharp contrast between their private opinion and facts, and you can always muddle the waters and spin the arguments to try and present why your opinion is "objective truth" and anybody who thinks otherwise is just "truth denier".

As a political propaganda, it's fine - everybody tried to prove their opinion is correct and the other guy's opinion is wrong. But it takes special kind of hutzpah to call one's opinion "objective fact" and try to paint all the other opinions as lying. Exactly the kind of hutzpah the "fact checker" sites are swimming in.

I think we are agreement. I am horrified by the recent demonetization of science, expertise, and investigative journalism.

My point is not that fake news isn't real. It is that when Gmail marks a weird Nigerian prince style scam email as spam, there is nothing controversial to the user about that. The user and the spam filter share the same opinion of what is spam.

The problem with fake news is that many conservatives disagree with tech companies on what fake news means. So if a user wants to share a BS info wars article, and is told he or she can't because it's a load of crock, they are not going to like that. People don't use social media to be told they are wrong. From the perspective of these kinds of conservative users, the spam equivalent would be Gmail marking an email from your mother as spam.

It's a hard problem because, unlike spam, it requires companies to disagree with their users.

Do you know they do independent verification of stories?

I have met lots of journalists. I have been interviewed for stories (on Bitcoin and blockchain stuff) many times. I have never met, heard of or encountered anyone whose job was to independently check facts or talking points provided by sources.

I don't have first-person experience with this, but I have heard about people being interviewed for articles in New Yorker, and they have been contacted by exactly those people - whose job was to verify the facts in the article (different from article writer, editor, etc.). I am not sure how widespread this practice is, but these people certainly do exist.

>Just look at the attacks on Politifact, Factcheck.org, etc... for people who don't think objective truth is a factor in "Fake news".

More likely explanation: people mistake personal bias for "common sense" or some kind of factual information, and then attack what they assume is "fake news" based on these faulty starting points.

> If you asked users to mark stories they considered fake, some would mark info wars and some would mark huff post.

It's possible that's the right answer, and that there should be more scepticism of stories from both of those sources.

Fake news didn't start online, though. There have been many paper publications which lie outside of mainstream scientific and ideological orthodoxy, e.g: Fortean Times, Daily Worker, Sunday Sport, which today we might call "fake news". Some of these had quite wide circulations and were available in many shops, and yet didn't precipitate any kind of crisis.

It would seem that fake news is perhaps a symptom of something deeper. It's possible that the mainstream press has lost the authority to suppress the fringes as public figures have learned to control their image in the media more (i.e: spin), presenting at least the appearance of collusion.

EDIT: It would appear that Joseph Pulitzer, creator of the prestigious journalism prize, was deeply involved in the rise of "yellow journalism", the "fake news" of the 1890s. Perhaps in a century from now, journalists will covet the Bannon Prize?

I did the fakenewschallenge, and my day job is related.

I'm not arguing it is easy, but there are plenty of things which make it worse than it has to be.

But, my understanding of things like the fake news challenge is that their definition of fake news is probrably news that has been shown by experts, scientists, video evidence, scholars, actual investigative journalists, or similar to be demonstrably false.

The alt-right is already aware that scholars, scientists, and generally educated people believe their news sources to be bunk and that has not changed their opinion. So even if facebook or whoever makes perfect software that blocks crazy info wars articles or other news meeting the definition above, conservatives will cry censorship and find another platform because their chosen platform fundamentally disagrees with them on what is true. People don’t use Facebook to be told they are wrong.

> Deliberate design choices are what made fake news effective.

These choices are made by whoever makes the fake news site. They are (ab)users of technology and not its creators. It's like saying encyclopedias are evil because you can take a heavy tome of encyclopedia and hit somebody in the head with it.

I am not saying solving fake news problem (if it exists) would be a bad thing. I am saying just like email is not evil because spammers exist, no technology is evil because fake news exist.