Facebook is harming our democracy
vox.com
vox.com
2) The article takes an interesting attempt at being even-handed: "it's not just right-wing Trump supporters who are being evil, but also left-wing Bernie supporters!"
3) The solution that Vox is suggesting is for Facebook to appoint a high council of editors who approve which stories are worthy of being allowed to be shared, and which are evil and duplicitous. For some reason I strongly suspect that the set of stories they want deemed good are those that align with Vox's editorial philosophy.
4) I'm curious how they would handle something like the Iraq war, where the two legitimate sides promoted by serious, responsible folks like Ezra Klein and the NYTimes were liberals who supported a humanitarian war and conservatives who supported a pre-emptive defensive war. How likely is it that the people who should have been banned from public Facebook discourse are those conspiracy nuts who thought the USA should stay out of Iraq?
Not "evil and duplicitous", but false and sensationalistic. I expect that Vox's opinion of what counts as both is biased, and you might think that Vox's own articles tend to fall into those categories. Certainly it is not always possible to assess truth in politics in a completely objective manner; I frequently read PolitiFact, and while I think they're decent, their fact checks often boil down to judging how accurately a short and necessarily imprecise statement describes a complex reality, leading to subjective rulings I sometimes disagree with. But some stories are just complete nonsense, like the "oversampling" example from the Vox article, and any reasonable observer really ought to agree on that...
Vox here seems to think that right-wingers generating click-baity bullshit is the big threat to democracy to design against. I agree that Breitbart et al. generate plenty more than their fair share of transparent bullshit. But I need strong evidence that that's actually a significant threat to the world.
What I see as the real threat is a cabal of the powerful defining strict limits on discourse to help generate consent around decisions that are toxic to most of the public and beneficial only to a small segment of it.
We've seen the latter happen recently: the Iraq War. You can get me to agree that the extreme nutty right believes lots of nutty, false things, but I don't see evidence that it has killed nearly as many people. Maybe if somehow Trump wins, and can attribute his win to false right wing bullshit, and he's actually as bad as the MSM says, it would work as evidence that that's a realistic threat model.
Then again, half a million dead bodies is a hard mountain to climb.
I'd be happier with it than human editors promoting Vox-approved stories, because 1) at least it'd be transparent what FB was going for and 2) it would be more fault-tolerant to an individual editor going off the reservation, by hook or by crook.
But regarding your main point, about using algorithms rather than humans: it might work to some extent against clickbait or low-quality articles, which is part of what the Vox article is complaining about, but another part is false information, which isn't always possible for a human to distinguish at first glance, let alone a computer. It often requires a bit of research or context. 'Real' news organizations often rely on asking experts in the subject at hand, who can contextualize the information much better than any computer; Facebook's hypothetical editors might not be expected to do that, but they could at least do some cursory Googling, which it's still pretty unlikely a computer could do effectively.
Yeah I had the same reaction. It was also somewhat rich that they cited a BuzzFeed article as evidence people are gaming our predilections for clickbait. Though, now that I think about it since it is clickbait, BuzzFeed would certainly be an expert there. And in a way, Vox also takes advantage of it. Here's a short selection of my favorite "Vox [explains x] in [<=500 words]" pieces.
http://www.vox.com/2016/8/29/12692546/obamacare-whats-wrong
http://www.vox.com/2016/7/15/12204172/turkey-coup-erdogan-mi...
http://www.vox.com/science-and-health/2016/10/4/13155916/201...
http://www.vox.com/2016/9/7/12817566/donald-trump-immigratio...
Last one for kicks. To be fair to Vox, they do produce some long-form stuff and often link it at the bottom of these shorter pieces, but these explainers are not helpful and, I think, produce a rather shallow understanding of extremely complex issues.
Unfortunately, that'd sort of undermine their entire business model.
A single humanbeing cannot possibly digest long articles on every newsworthy topic. Therefore the short form is an important staple of news coverage. Of course you will be left with a "rather shallow understanding of extremely complex issues", but the point is that you are made aware that the issue exists at all.
Interesting 'fact' to cite. How many emails did that server send again?
But they should not censor posts under any circumstances, even that one. The problem is the way stories trend and the order in which the feed is presented. The current algorithms and 'like' counting maximize shocking statements that confirm world views exactly because that will maximize user retention. Accuracy is completely deprecated.
Facebook knows this and is well aware that nudity (let alone porn) would overwhelm their site because it would be overwhelmingly favored by their systemic selection process. So rather than fix the process they censor that. This leaves things like accusing people of satanism or other false inciting statements to bubble up instead.
Unfortunately, facebook has no fiscal incentive to fix this. And it may not be a coincidence that UKIP, Trump, AfD, Le Pen expanded greatly after the emergence of facebook.
The reason I count the story as false is that all the available evidence points to it being an ordinary mass mailing system: the horde of similar domains owned by Cendyn for the other hotels they do business with, the e-mails from it that have been found, the long trail of evidence showing Cendyn have been in this business for two decades and have had the Trump Organisation as a customer for 9 years, the existence of numerous DNS lookups not in the original logs offered by those pushing the story... When the only evidence is misinterpreting ordinary, boring activities, then it's a conspiracy theory. This is why it hung around for months without any of the publications that were offered it taking the story.
I'm not a fan of click-baity articles, but the chance for all voices to at least be heard trumps (no pun intended) relying on what we may dub the "media elite" for our news. Time and time again they've shown they aren't the objective neutral actors in ALL cases they claim to be. Of course the WSJ or NYT will have a better overall track record than say buzzfeed, but that doesn't mean there won't be cases in which the the aforementioned elite media organizations don't try to skew the conversation.
A good example of this is the ND pipeline. None of the established players are giving it the coverage it would seemingly deserve, without YouTube, and smaller outlets the majority of the country wouldn't know anything is happening.
We fund content distribution (Facebook) and content creation (actual articles) based on how many people are looking at them, so journalists produce clickbait and Facebook gets you addicted to it. We can call it "maximizing daily active users" instead of addiction if you like, but it all sounds the same to me. If you're asking Facebook or journalists to make less money to create a better world, don't be surprised when they say no.
Instead, individuals must take on the responsibility to fund the production and distribution of content they find valuable. That won't eliminate bias, but it'll eliminate the very strong financial incentive embedded in capitalism to create a compelling alternate reality for each individual on the planet. It is tearing us apart. Start a recurring donation to a nonprofit news organization today, then turn on your ad blockers. Not just for intrusive ads. For everything.
http://www.theatlantic.com/technology/archive/2014/09/facebo...
If not, why is 50℅ somehow bad while 100℅ is ok?
Newspapers are private companies too.
I'm not sure that's true. In a 2016 article Pew Research has TV as easily the number one source. http://www.journalism.org/2016/07/07/pathways-to-news/
I'd be more worried by Fox News than Facebook. Also Breitbart and the Daily Stormer can be scary experiences.
Edit: oooh hit a nerve there.
When everyone is upset and there are many points of view there is no common way to move forward but there are many hot heads that are willing to shoot first and ask questions later. Imagine a million hot heads without a common goal but the willingness to fight and we can see chaos without results.
People get upset at the "do nothing congress" because they can't get X done but people aren't willing to admit that the reason is that voters have sent individuals with very diverse ideas to try to get things done. Voters are the ones that are pushing them to not compromise any idea or be punished by being voted out. I can see a future where one person that can use social media very well can push people to vote in ways that we consider distasteful now. What will happen then? Groups will erupt with opposing view and many will be ready to fight.
We've heard allegations that the voting system is rigged but that's very unlikely. We have laws and watch dogs that prevent that in any significant way. We don't have the same for social media but we know that it's possible to manipulate it, even by foreign powers, and that's not illegal worse yet it's hard to impossible to prevent. Yet, any thing published whether true or not becomes true by effect if enough people repeat it. It's hard to even contemplate how that effects a democratic system.
The founding father created a representative government because they knew that rule by majority can be as distasteful as government by a monarchy or emperor-ship. They thought a functioning government needs representatives that can sort out what's needed for a better nation. With everyone having a voice that's going to get extremely difficult. Social media is about to let the US test out its governmental system, lets hope it can pass the trouble ahead. Can the US stay together as a nation when there is no common ground among the citizenship?
One week ago, FB blocked an event page of Independence March, which traditionally takes place on November 11th (Polish Independence Day) in Warsaw. This is the biggest mass event of Independence Day, having more than 100k participants each year.
FB also blocked or removed pages of NGOs and political parties, which are organizing or support the march (some of them had 80k or 170k followers) and personal accounts of people involved in these organizations.
Then FB went full rage and started to block personal accounts of everyone who invited or even positively mentioned the Independence March, including for example the personal account of editor-in-chief of the second largest daily newspaper in Poland (https://twitter.com/sjastrzebowski/status/793001362070052864).
The most extreme case was the personal account of a MP, who wrote on his timeline: "I will be [on the Independence March] along with my family, whether FB likes that or not." (https://twitter.com/jakubiak_marek/status/793497135954202625...) His profile was blocked for 24 hours after that.
Another case was a personal profile of a retired Intelligence Agency officer, who revealed in a FB post, that a local coordinator of an anti-government liberal-left protest movement during the communist period was a colonel of Soviet-dependent military intelligence agency.
All of this happened just within the last month. FB actions generated a huge pushback and hit the headlines. Deputy Minister of Justice qualified FB actions as "censorship". Minister of Digitization tweeted that she "asked FB management for a talk". Many people started deleting their FB accounts in protest.
FB got frightened and reactivated the event page of Independence March, but many nationalist/conservative organizations profiles still remain blocked.
Humans and algorithms will have bias, but a diversity of bias is good.
If industry cannot implement some self regulation, it risks finding itself with external policy and regulatory controls.
News networks were much worse, they could arbitrarily censor anything that did not result in a higher TV rating, more advertising revenue, or that didn't represent the interests of the investors/donors/advertisers... oh, wait. Same on Facebook to some extent :-)
That's why it's important to decentralize the Internet.
Doesn't censor opinion I don't like - Harming our democracy.
Eg:
Trump is harming our democracy .
Wikileaks is harming our democracy.
FBI is harming our democracy.
Pepe the frog is harming our democracy.
Russia is harming our democracy.
That is a point the article makes, though. The algorithmic approach is effectively equivalent to particular group of people doing the approval (or rejection). What group that is manifests itself via the way the algorithms are coded and the data that are fed into it.
Edit: This is similar to the conclusion people are coming to regarding algorithms for court sentencing[0] and other areas where algorithms are becoming increasingly used. It's dangerous to conclude that an algorithm is better just by virtue of it being code. That code (and/or its training data) may be reinforcing existing stereotypes and biases.
[0] https://www.engadget.com/2016/06/26/wisconsin-sentencing-alg...
It's similar to a hiring process: which is better, having a known procedure and rubric for choosing which candidates to hire, or just allow a manager to go with their gut and choose their nephew? There's a sense in which they are the same--at some level, designing a system must still take into account human values and biases. But that's still a huge step forward from having people, even smart people, make gut decisions just based on what's in their head, even if that's still a decision driven by human values and biases.
I don't see how that follows. The biases of an algorithm or beurocracy aren't inherently superior, they are just inherently repeatable.
The conclusion there is simply a theorem: it's mathematically impossible to be both well calibrated (a black and white person receiving the same risk score have the same probability of committing crime) and also racially balanced (similar levels of false positives), except in trivial and unrealistic cases.
https://arxiv.org/pdf/1609.05807v1.pdf
See also this WaPo article explaining in simpler terms: https://www.washingtonpost.com/news/monkey-cage/wp/2016/10/1...
But the thing is, this theorem applies to any decision process - human or algorithm. You can't escape mathematical impossibility results just by using humans (e.g. NP Complete problems are still hard even if done by humans).
Human processes just add additional bias, e.g. the editorial slant that Vox wants Facebook to add, or the (alleged) bias that Facebook's human editors added before Facebook went all algorithmic.
Bias with respect to what? As you say, there is already bias baked into the data collection and the algorithmic choices.
The bias that human editors introduce is different, but not necessarily larger, however you even measure it. There are also myriad human choices behind the choice and deployment details of the algorithm.
An important plus for human editors is greater interpretability and greater transparency regarding the biases the system ends up showing.
Machines simply don't do this.
As you say, there is already bias baked into the data collection and the algorithmic choices.
That's not what I said.
What I said is that you can't have collective equality (e.g. same rate of false positives, lack of disparate impact) and also accuracy (getting the right answer) except in trivial/unrealistic cases.
Human editors are fundamentally less interpretable and transparent than machines. You can easily interrogate machines and test for bias; how do you do that to humans?
Or, to take a historical example, why did colleges switch from algorithms to humans when the supreme court said that transparent racial bias is forbidden?
Okay, that may have not been the best example to reference.
> Human processes just add additional bias, e.g. the editorial slant that Vox wants Facebook to add, or the (alleged) bias that Facebook's human editors added before Facebook went all algorithmic.
But how you can craft an algorithm that takes human processes (clickbait, investigative journalism, breaking news, photos of cats) and provides an output free of biases? Given the fluid nature of the input biases, is it even possible to craft an algorithm that accounts for those biases without adding in (or substituting) other biases? What is the "objective standard" by which you measure and correct the input biases? Just because there's an algorithm doesn't mean the output is bias-free.
Historically, that's exactly what newspaper editors did. The algorithmic approach has the potential to offer greater transparency, but not if the algorithm isn't published.
Furthermore, now that Facebook is in a near-monopolistic position when it comes to picking winners and losers in what we read, acknowledging that fact publicly would be the bare-minimum first step they could take towards accountability. Then we can have a more open discussion about how news stories ought to be judged and propagated (or not).
This isn't really true. There's actually a mathematical discipline devoted to treating statistical algorithms as black boxes, and assessing their accuracy/generalizability based solely on test data/broad intrinsic properties of the algorithm (e.g. continuity parameters).
It's called "machine learning".
From what I can tell, the fundamental distinction between machine learning and statistics is that machine learning provides mathematical guarantees based solely on black box testing of algorithms and very broad properties of the distribution/algo. In contrast, statistics cares about the actual underlying process.
But a human editor, on the other hand, is much less transparent -- basically, the best justification they can offer is "trust me, I know what I'm doing".