Facebook says it was wrong to ban Canadian MP’s Christmas message

Companies like Facebook & Google are able to scale up in part because they're avoiding spending money and time on moderation & customer support. If they were legally required to maintain a certain level of support, then that would help curtail their speed of scaling up, and probably be a good thing in general.

Comparing customer service of Facebook to a mortgage servicer is tricky because of the value and the volume. Facebook has orders of magnitude (I'll get 10,000x more) content to moderate than a mortgage servicer where they really only answer questions when something goes wrong (3x every 5 years) the stakes are a million times higher.

The options for social media are really free (or cheap) to use and occasionally bad moderation or pay $5 per post or comment ($10 for risky ones) that see multiple human moderators to limit abuse and false positives. Or no social media. My point is the amount and quality of moderation people say they want is more expensive than they'd pay for these services, but there's still a lot of value to them, so we're left with imperfect moderation.

This is typically discussed in terms of:

- Avoided negative externalities. FB/Google get to ignore the costs of insufficient moderation by imposing those on their users or third parties (say, the victims of the Myanmar genocide).

- Monopoly rents. By securing a position in which they are the arbiters of access control, these giants can set both prices and terms for participation.

There's another element you're suggesting to me as well, which I don't think I've seen anyone articulate before: large monopoly players such as Facebook and Google effectively emerge into a "naive environment", one in which the scale which they provide has not yet evolved threats. Any threats which do emerge co-evolve with the monopoly.

Subsequent entrants don't have this advantage. Whilst they do have the benefit of techniques for combatting such threats, they must address the threats from the beginning.

The threats might variously be spammers, various network attacks (most especially DDoS), sockpuppet and other misrepresentation and manipulation attacks, harrassment, and the usual laundry list of Trust & Safety issues.

The notion of co-evolving threats comes from biology generally, though I'd strongly recommend Kyle Harper's The Fate of Rome which postulates that civilisations (themselves social networks) and their plagues co-evolve alongside one another. I've just learned that the Santa Fe Institute crowd are fans of Harpers (Geoffrey West and David Krakauer, in a recent Complexity podcast episode), as well as Neal Stephenson.

The news there is that the system is so bad that it has a very high false-positive rate, including something that should never have been a false positive in any proper system. That's like the police arresting the mayor of San-Francisco for a robbery that happened at the same time in Miami, and was perpetrated by a male. Surely, they let her go after the lawyers got involved, but wouldn't you call a policeman who makes so obviously wrong arrest an idiot unfit for the job, regardless of the fact that the error was later corrected? Wouldn't you call for re-training, or maybe even firing such a policeman, even if you agree that there should be police in general and police work is important?

I think if we design artificial idiots, and entrust them to gatekeep our society's communications, it's newsworthy to point that fact out.

While I'd love to be optimistic, not all of us are Canadian members of parliament.

Imagine if this guy wasn't a Canadian MP, and instead just some random person without reach or influence. Do you think this would still have gotten noticed solved as quickly as it did?

Stuff like this always reminds me about this story [0], that became only a story because the editor of a rather well known Norwegian newspaper ran into that problem himself.

There is no telling how many people before him ran into the same issue, and just went; "Well, I guess then I can't do anything about it" because they didn't have the ability to publish this issue in a renowned traditional news media, which actually gets it attention.

At best those other people might have described their experience somewhere, and probably got called conspiracy nuts for it.

[0] https://www.bbc.com/news/technology-37318031

Time and time again "crazy" conspiracy theorists end up being right.

In what way was the incorrect algorithmic take-down of the MP's Christmas message due to the size of FB?

and then the broken up parts just collaborate with each other and you get the same result with some extra paperwork

It's possible that content-based suppression happens, but it's also possible that the genericized phrase organically generates more engagement by virtue of it being likeable to way more people. If all your followers celebrate the same holiday then it should be a wash but if huge chunks don't then it's purely the numbers doing their thing. Even some who celebrate the mentioned holiday might find the message exclusionary enough to not engage (a woke feedback loop, you could say), further burying it.

Do you have some sources for further reading? That sounds disturbingly Orwellian.

[citation needed]

There's a fairly rational psychological explanation to it. When people cannot achieve a certain goal for a while, they convince themselves that the goal is not worth it, or tainted in some way.

Christmas has been traditionally associated with family, kids, happiness, seeing retired grandparents, etc. All these things have been sliding out of reach for the middle class with no reversal in sight. So this creates tension against people that can still afford it, and a perception that these goals are unfair and should not be pursued.

This is to be expected - it's not like the journalists writing the politically correct articles, HR people defining acceptable language, or moderators deciding on what gets approved are paid enough to afford what used to be a regular middle-class life 2 decades ago. So they rebel against it by trying to stigmatize it in the way they can.

And it's not like you can blame them either - the human brain is very adaptive at picking goals and means. If you take personal prosperity through hard work out of question, people will find other goals to pursue, it's just they will be much more divisive...

This year, Merry Christmas and Happy Hanukkah are almost a month apart (Hanukkah started on Nov 28th). So you'd have to either have Merry Christmas in November, or do it twice anyway :)

> FB biases are too extreme. The best thing for FB is to be broken up.

That's a non sequitur. Breaking up FB will not de-bias "the algorithm". We'd have one company doing VR, one doing phone calls, one doing image sharing, and finally a company running the old social network with the exact same recommendation and content-moderation system.

I share a similar sentiment/observation. I am OK with the world where the occasional mistake occurs with algorithms. But when things like this continually happen with Facebook algorithm, it feels as though they are trying to outsource too much but should involve humans to computers.

In short, their tools tend to be overly restrictive, airing on the side of rejecting. I’m not sure thats right

Facebook sucks and I never use them. That being said, Facebook moderators who manually review reports are routinely traumatized and diagnosed with PTSD, and Facebook catches fire for that.

So I don't see how this can go both ways. How can we ask Facebook to automate this system and flame them for it not being thorough enough, and then simultaneously flame them for implementing the thorough system of human review? As much as the company sucks, this seems like a no win scenario that isn't fair even to a reasonable company.

This is one instance of someone with a clear political affiliation; the question is whether it would happen to someone on the other side, or with little discernable political affiliation.

> our God and Savior *your God and Savior

On this day, Thor's Day, I wish everyone here a "Merry Christmas" celebrating the Yule season. How appropriate to have our saturnalia on Saturn's Day this year!

Please don't post unsubstantive and/or flamebait comments. It's not what this site is for, and it destroys what it is for.

https://news.ycombinator.com/newsguidelines.html

Edit: it looks like your account has mostly been doing that. We ban such accounts, so if you would please review the guidelines and fix this, that would be good.

Never assume malice when machine learning will suffice as an explanation.

"Just don't use it" doesn't apply to things with a network effect.

> You do not have to use FB/Meta.

Unfortunately there is a so-called "network effect", which prevents quitting for many users.

There are laws that do impose obligations on some privately owned systems regarding speech they must or must not carry. Being "privately owned" is not an exception to these laws in the US, let alone other jurisdictions.

Regardless, rightful censorship is still accurately described as such.

And, this doesn't seem to be a story about censorship, IMO, because it was clearly accidental. It's a story about bad ML.

Yeah, but this narrative is just as wrong and tired as the 'censorship' one. Companies are not obligated to carry your speech on their platforms, but we can and should be able to have a conversation about whether or not their methods of moderation are sound / reasonable.

It clearly is censorship, it just happens to be legal censorship.

1) It's still censorship, even if it's not government censorship. The word applies.

2) It's perfectly fair, and a good idea, to call out a thing as bad even if it's legal.

The job of roman Censors was to supervise public morality. The term censorship is appropriate

>"FB/Meta has no obligation to carry your speech"

Why not? I believe it should, given its size. The phone company has an obligation to carry my speech. Net Neutrality would require ISPs to carry speech. There are a whole host of things the government can compel "privately owned" systems/companies to do.

In these kinds of discussions, scale matters. There is a big difference between compelling a company with a billion+ users to act as a platform and forcing a small business to do the same.

To be pedantic, Facebook definitely does have the capability of censorship:

> Censorship, the suppression of words, images, or ideas that are "offensive," happens whenever some people succeed in imposing their personal political or moral values on others. Censorship can be carried out by the government as well as private pressure groups. Censorship by the government is unconstitutional.[1]

Facebook has enough power over communication at large to effectively suppress speech should they choose. Censorship does not have to be perpetrated by governments. Hence the terms self-censorship and corporate-censorship[2].

What people really mean when they talk about censorship is first amendment rights. Here, you are correct. The first amendment protects against government censorship only - nothing Facebook does could infringe your first amendment rights.

[1] https://www.aclu.org/other/what-censorship

[2] https://en.wikipedia.org/wiki/Corporate_censorship

In the 1960s, certain restaurants and other establishments argued that they were private, and could choose who was allowed to enter and do business there.

They were making the same mistake you are - arguing that they were not obligated to be even handed and non-discriminatory in the services they offered.

And you don't have to read our comments where we share our anger at Facebook/Meta.

Surely you don't think corporations can't be criticised?

It is censorship, though. Not government censorship, but don't you think that one day in the future, once these non-government censorship systems are in place, that a corrupt or evil government could abuse those systems and make them defacto government censorship systems (we're already partway down the slope with things like CSAM and copyright monitoring, although the former is especially heinous and so censorship of that type of content is desired).

Not every provider of user-generated content should have to abide by these rules, because the damage that can be done by a relatively small forum or website is usually minimal (although see sites like Kiwi Farms that may not technically directly encourage harassment but where it is nonetheless often perpetrated by sadistic individuals who participate in their forums), whereas the damage that can be done by a Facebook or a Google is much, much greater.

>Censorship is the suppression of speech, public communication, or other information. This may be done on the basis that such material is considered objectionable, harmful, sensitive, or "inconvenient". Censorship can be conducted by governments, private institutions and other controlling bodies. Governments and private organizations may engage in censorship.

There is nothing that presupposes that the entity has the obligation to carry your message. Just that the entity removes your message on the grounds that they do not like the material.

This comes up every time someone mentions that X(a private party) censors something. Yes, X is not the government. Yes, it is still censorship. If undertaken by the entity on it's own, it may not be illegal. If the entity had government influence in it's decision, than it may be.

There are some here that argue that the status quo is fine and others that argue that we need to change our laws and regulations to protect individuals from more privet enterprise censorship. That is where the debate is, not whether or not censorship exists.

> The 'censorship' narrative is wrong and tired. FB/Meta has no obligation to carry your speech.

they should when they become a mega-corp