This is a fair point, but I think the political aspect is an "emergent property" if you will, rather than an inherent one.
If Facebook limits the filtering to engagement, then it isn't the fault of Facebook that political content is engaging. That's just human nature. Disasters, outrage, politics, polarizing topics - these are all popular topics both online and off-line, and apread quickly as town gossip well before Facebook.
It is only when Facebook steps in and says that particular topics need to be exceptions to the filtering rules that apply to everything else, that they make themselves into a political actor.
For instance, let's say that the news feed showed you content based purely on number of likes. If political posts get lots of likes that isn't Facebook's problem. If the same ranking rules apply to all posts (# of likes) then they would remain neutral. As soon as Facebook says "content from x person will have their ranking artificially changed to reduce/increase engagement with it" thereby making an exception to the rule that applies to everything else, they have now become a political actor.
How could it not be the fault of Facebook when Facebook designed the algorithms that are creating all of the divisiveness on Facebook?
If I build a bridge intending it to stay up and it happens to fall down 6 months later, I'm responsible for it. Facebook created an algorithm that divides people politically and that surfaces content that is provably fictional. So they should be held responsible for it regardless of their intent. They don't get to invoke "common carrier" status when they're writing software that makes decisions about what you do or don't see. What makes a telephone a "common carrier" is the fact that the telephone doesn't decide who you call.
It doesn't matter whether it's software or a human. What matters is that decisions are being made by Facebook about what you do or don't see.
Whether or not it is intentional is immaterial to the effect. The law doesn't care about your intent. I wouldn't intentionally dump toxic waste into a river but I'm liable for dumping whether I intended to or not. Mark Zuckerberg can't just throw up his hands and go "oops it's software I can't help it" when it's his company that made all of the decisions about how the software works.
This isn't correct. The law in most modern democracies, as far as I'm aware, is very concerned with intent.
This why we generally define murder and manslaughter as distinct.
Murder is the unlawful killing of another human without justification or valid excuse, especially the unlawful killing of another human with malice aforethought.
Truth is not the point of social media. Facebook isn't an encyclopaedia. Humans already gravitate towards groups that validate their opinions. It makes sense for Facebook to show people content they want to see. It is incredibly Orwellian to say "Facebook should only show people content which corrects their wrong views". Facebook isn't a social conditioning tool. I find the alternative to "misinformation" much scarier. Misinformation and being mistaken are human flaws we will always have, and therefore any social groups will have them by default. Using technology as a tool to condition people out of their views against their will is scary.
The information is out there. There are reliable news sources. There are reliable databases and encyclopaedias and journalism. If people choose not to read them then that's on them.
>It makes sense for Facebook to show people content they want to see.
The problem is, Facebook doesn’t show people content that they want to see. They show people content that they will engage with. That’s a very important distinction.
HN algorithm/moderators actually explicitly do the opposite: if a thread gets too many comments too quickly, it’s ranked downward. The assumption is too many comments too quickly indicates a flamewar and the HN moderators want to keep a civil discussion. The approach Facebook takes is to “foster active discussion” which on the Internet typically means a flamewar. Noting generates engagement like controversial political views. So that’s what Facebook’s algorithm/moderators show to their users.
Facebook absolutely is a social conditioning tool, it’s designed from the ground up to show people content that stirs their emotions enough to click “like” or the mad face icon or even leave a comment and wait around until someone replies back.
My point is that this what happens in real life. People will continue to engage in stuff that they want to engage in. Facebook doesn't force people to engage in anything.
I think it is far worse to attempt to condition people by showing them things *they wouldn't otherwise engage with". What is scarier, showing somebody something they want to engage in based on their past behaviour, or showing somebody something they wouldn't have otherwise seen if you hadn't gone out of your way to shove it down their throat?
People seem to want Facebook to make people more placid. Oh you have extreme views? Here's, let's condition that out of you by only showing you more moderate stuff. Oh you think x is bad? Let's not show you anything to do with x so that you'll hopefully forget about it and not engage with that part of your brain any more.
Like I've already said, this alternative is far more Orwellian and far more of a tool for social control, than simply optimising for engagement.
> I think it is far worse to attempt to condition people by showing them things they wouldn't otherwise engage with". What is scarier, showing somebody something they want to engage in based on their past behaviour, or showing somebody something they wouldn't have otherwise seen if you hadn't gone out of your way to shove it down their throat?*
I don't think that makes sense, and I don't think that's what anyone's advocating for.
If you friend someone, or follow a page, or whatever, you are explicitly saying "I want to hear what this person/group has to say". They aren't saying "I want FB to carefully curate what this person/group says in order to increase my engagement of FB". FB shouldn't promote, hide, or reorder anything coming from someone who I've explicitly chosen to follow. It should just show me all of it, and let me decide what I do and don't want to see.
That's no distinction at all, what people engage with is just one effective way of measuring what people want to see. HN simply optimizes for something else, that's no less of a social conditioning tool than optimizing for engagement, just in a different direction. You could say that it's designed from the ground up to show people content that stirs their curiosity enough to comment cautiously, or to hide content that stirs their emotions enough to engage strongly.
Facebook has a fact-checking program. That program has third-party fact-checkers. Facebook has been documented as pressuring those third-party fact-checkers to change their rulings.
They can't impute that they checked facts, remove postsings they believe are incorrect, and then quietly put pressure on the fact-checkers to have a different "opinion" as to what is "factual".
I agree with what you're saying; pointing a finger at one particular thing and saying "this gets suppressed" is a bad solution. In fact it probably opens the door to a whole new category of problems.
But... I think what we're seeing with political content is just a symptom of the real problem.
> Disasters, outrage, politics, polarizing topics - these are all popular topics both online and off-line, and spread quickly as town gossip well before Facebook.
This is true. But when information spreads through people's conversations with each other there's limits to how fast it spreads. There's also a lot of room for dialogue and different perspectives. If I have some silly conspiracy theory that I want to spread around, it's going to be pretty hard to convince the people around me that 5G is going to activate microchips that were injected into my bloodstream. They will likely point out that basic laws of physics don't really allow for that. But if I know how to game a social media algorithm[0] to connect me with millions of people that are susceptible to that kind of thinking, I could convince a shockingly huge number of them to believe it[1]. Especially if the social media platform isolates those people from opposing opinions and connects them with people that think similarly.
I think social media is like removing the control rods from a reactor. Those basic human flaws are now being amplified and capitalized on at a scale we can barely even grasp. And it really doesn't matter if Facebook, Twitter, etc. are "at fault" or not. It's a fundamental problem with this services and the problems will continue to get worse.
Whether or not a property of a system is emergent does not seem particularly relevant to discussing its effects. It would be nice if we could just look at the way a platform like Facebook works, or see that it does not intend to be a "political actor", but the reality is that its effects are destructive. It has a strong effect on politics. The effects of Facebook are not outside of its responsibility simply because it isn't programmed to be political.
> let's say that the news feed showed you content based purely on number of likes
Does any site actually do this successfully? It seems to me that even sites that lean heavily towards algorithmic curation (including HN) still have an element of human veto.
it's not emergent, it's inherent. facebook is political because it was designed from the start to control information. that's coercive, designed literally to change people's behavior, whether covertly or overtly, which is the essence of power and the political. facebook employees cannot absolve themselves of responsibility by saying it's the machine that did it, or that other people were involved and complicit.
being political is not an incidental facet of facebook, it's a core intention.
1. This sounds like a somewhat naive take on filtering algorithms.
2. Does it matter if it’s Facebooks “fault” or not? The issue is their power.
Imho ideally they would acknowledge and accept responsibility for their power and in the US at least there would also be some laws regulating them in this regard.
No it isn't. The creator chooses which data to feed in, and which not to. Garbage in, garbage out. If Facebook positions itself as the arbitor of data and actively stops people from posting, then it must take responsibility for what is on the platform. If it simply allows people to sort the data that anybody can post, then any properties of that sorting (whether by likes, or date posted) are emergent.
If Facebook limits the filtering to engagement, then it isn't the fault of Facebook that political content is engaging. That's just human nature. Disasters, outrage, politics, polarizing topics - these are all popular topics both online and off-line, and apread quickly as town gossip well before Facebook.
It is only when Facebook steps in and says that particular topics need to be exceptions to the filtering rules that apply to everything else, that they make themselves into a political actor.
For instance, let's say that the news feed showed you content based purely on number of likes. If political posts get lots of likes that isn't Facebook's problem. If the same ranking rules apply to all posts (# of likes) then they would remain neutral. As soon as Facebook says "content from x person will have their ranking artificially changed to reduce/increase engagement with it" thereby making an exception to the rule that applies to everything else, they have now become a political actor.