Leaving aside how incredibly illegal your proposal is, it's also unworkable and would make the internet worse if even attempted.

If I have to communicate to every bad actor how I identified them as a bad actor, that only helps them sidestep my prevention mechanisms.

I'm sure it can be frustrating to face a blank wall, but fortunately, Facebook is a waste of time that nobody should be using anyway.

1. User gets banned by Facebook without an explanation

2. User asks Facebook for an explanation

3. Facebook declines to answer

4. User goes to the government regulatory body with a complaint

5. Regulatory body forces Facebook to provide a reason

6. Facebook tells user what they violated and when

Step #5 would allow the regulatory body to require identification to make sure that the user is under their jurisdiction (or you just have to show up in person). This would allow them to identify bad actors and would decrease the chance that the system could be abused.

Obviously this should only apply to platforms/services that are beyond a certain size (>1M users?). Also, I wouldn't be surprised if some consumer protection agency doesn't already have powers akin to this.

Useful/good idea is one for the debate floor. I imagine in this case, for example, Facebook would send Mr. Birch an email telling him that he distributed nude photos via Messenger and they have a 0 tolerance policy towards that. That doesn't really move the situation forward from what is already being publicly written about in an article. It increases the cost of competing with Facebook too, which is probably an own goal given the outcomes the regulation is trying to achieve.

There are a lot of things in life that do not need a legislative response. Being banned from Facebook is one of them.

In short, you are proposing that we penalize good actors for actions of bad actors for no other reason that it creates inconvenience for a platform. That is not acceptable.

How you differentiate, I dunno. That's the problem.

Would you argue that a person accused of crime X should be told nothing about the nature of their transgression? To me it sounds very problematic.

I have been banned from a couple of forums/groups and I really have no clue about what might be the reasons but I don't insist there were none.

Why can't Facebook just have a hotline/chat that can verify a user's identity and tell them more, preferrable in less than 5 months of spamming their email?

And in the case of a permanent ban the user should at least be able to download all their account data; this feature is already implemented for compliance with GDPR anyway and the online platform wouldn't lose anything by at least not making pictures etc. inaccessible forever.

What actually seems going on here (as well as on YouTube with copyright stuff) is a wild misuse of AI technologies. Like if the police was giving people life sentences just because they had a weird accent or suspicious face triggering a neural network signal - perhaps deploying such measures on global scale could help catch a lot of murderers but undoubtedly it would also ruin many innocent people and this can hardly be considered OK.

So we don't even have to regulate Facebook in particular, just regulate large-scale AI usage.

It really depends on the level of detail you give, and how timely you are.

"You were banned for bot-like activity" and no further clarification at least lets the user know this was definitely in error, and gives actual spammers very little information.

"You were banned for bot-like activity due to low activity and engagement" gives spammers a bit more information, but only the really incompetent ones haven't figured that out by now - you might not want to advertise it for the first month or two, but real users might not realize they need to be fairly active after making an account (I'm from the older generation where we'd lurk for a while to observe social norms before commenting)

"Your account was banned for posting nudes / racism / etc." actively increases the effectiveness of those policies since now people will be more worried about edge cases and less willing to get near the topic.

Since you're using automated tools to catch people, you can also automatically communicate this. If you're really paranoid about bad actors, you could even hold on to that information and only include it in an automated email when someone contacts support (and then only replies to that first automated response actually go to your REAL support process)

If you don't, you're the bad actor. If specific actions lead to a ban, the way to sidestep that is to not do those actions. If you don't even look at the actions but some BS meta stuff, or have wishy-washy rules you enforce randomly, you're the bad actor.

It is often the reason for corporation like fb and google for closing account. It is totally BS. Spam is NOT a big deal compare to censorship. Any competent developer can use Bayesian algorithm to counter spam on a personal level, and the code can be just sitting on the client side. The big problem is these data company won't be able to get your data.

Not getting tech support lowers the profit margin for the big corp. However, automation and AI are NOT a replace for good customer support. Get some freaking real humane tech support in relation to the number of active user you have Google/FB. This is one of reason why the world hates you.

If an innocent gets banned and there's no explanation or way to reach anyone, you're Josef in The Trial.

Have programmers and marketing people in the Valley ever read any Kafka or Orwell?

Or they just ignore the likeness and go about as if nothing.

An oil refinery is able to make it another year because they dump all of their waste into the river behind the factory. Proper disposal is expensive and difficult. Is this fine?

There's a very valid argument that rules like this would just entrench existing players and hurt startup competition. I'm fine with this.

Running globally distributed, massively critical services like Facebook, Google, Microsoft, etc do, is a similarly massive responsibility. Parler incited an insurrection. Telegram is used to incite violence. Jeeze, plenty of this happens on Facebook & Twitter too. We're far, far past the golden days of scrappy startups deserving to be able to connect the whole planet without consequences.

The only reason why the oil refinery metaphor sounds so obvious is because we understand the consequences of dumping toxic waste. We understand them, today. We did not understand them 100 years ago. We lived through the consequences, and built laws to dampen them. The internet is, what, thirty years old? Its a baby. We don't understand its impact on society; we want to believe its positive, but we can't say for sure, and there's a ton of evidence pointing to it being just as negative as positive.

GDPR is a step in this direction; it does enable existing players to reinforce their moat against scrappy startups. Its especially harsh when one considers that it happened in Europe, which has almost no homegrown mega-tech sector like the US and China do. But, its still necessary. You're saying we should enable tech companies to act with essential impunity because it would be difficult for them to run their services in a way which is ethical and just. I refuse to sacrifice ethics on the altar of capitalism; I don't care how hard it is, and I don't care how much it costs them.

As for the legality: Many things are legal, or illegal, until they aren't. The funny thing about laws are, we make them. They're not etched into a big stone on a mountain, forever static.

Or you just use a wrong business model.

In the world of today I can see no reason why Facebook (I mean as a social network, not just as an advertisement and analytics store) could not be a paid service. People pay for Netflix anyway while they would probably choose Facebook if required to choose and keep access to just one of the 2 services.

I can see Facebooks side of this - when you give a specific reason, you open yourself up to a lot of counter-arguments, possibly outraging news articles for a bad formulation or a not fully specific reasoning and you give intel to spammers on how you detected them. Additionally, having service reps find and cite the exact violated terms costs a lot of time and it would not even matter - FB is free to ban you if it wants, the specific reasoning would be more of a FYI.

And, of course, in cases like this, where the ban was presumably unjustified or unnecessary, this is rather cruel. I fully agree. But on the other hand, FB bans probably hit bots 99% of the time and, even if not, are usually justified [0]. Opening up such a large attack surface and spending large amounts of money for the few cases where it hits a innocent (?) person is just not viable, and, as bad as it is, I can understand Facebook here.

The moral of the story is, as mentioned in the article, that the actual error was to rely on Facebook.

[0] I have obviously no hard data on this, but if this would happen a lot, it would be quite bad press. So I'm very sure that FB would tackle this problem if users would actually become scared of a false ban.

If the counter-arguments are reasonable, then Facebook _should_ be attacked.

> Additionally, having service reps find and cite the exact violated terms costs a lot of time and it would not even matter - FB is free to ban you if it wants, the specific reasoning would be more of a FYI.

I think the implicit implication is that Facebook should _not_ legally be allowed to ban you only if it wants. I would personally get behind such legislation. Sure if Facebook were just a social platform that's one thing, but it is tied to other services (e.g. Oculus) and is an identity provider for many other services. I don't believe Facebook should be able to have it both ways.

Yes. But this will probably end in a lot of arguing, which is good for spammers, offenders and innocent alike, but bad for Facebook in practically all cases - so seen from their side, it is all downsides, especially when the innocent make up a small proportion. Don't get me wrong - I don't think the result is good, but I can understand FB acting the way it does.

> I think the implicit implication is that Facebook should _not_ legally be allowed to ban you only if it wants. [...]

I'd agree to that, but I think we need to treat carefully to not create something wrong here. Forcing FB to allow you to have an account might lead to required identity documents across the industry, for example, which would hurt privacy even more.

OTOH, FB should not be able to simply "void" your hardware. Steam and the Nintendo Switch, for example, only ban you from multiplayer and social functions, but you can access anything you bought still (I think). That would be a reasonable solution for the Oculus as well. Generally, removing the need for an account from hardware by legislation, unless proven to be required, seems good.

Still, this would not help the guy in this article and I actually don't think we should force Facebook to allow anyone to ping random people or justify retracting free services. He should not have relied on Facebook as backup or sole point of contact. Yes, FBs conduct looks bad[0], but if he would've been realistic about the problems of a free services, the ban would be more of an inconvenience.

[0] I don't want to take sides here without knowing the details of the story.

Greatly doubt that given overt spam being so prevalent.

If, hypothetically, 99% of bots were banned and there's enough lingering, you would doubt it to be true.

Although, I tend to agree with you about spam prevalence. I've reported dozens of spam/scams/fake accounts and almost every time, I'm told that the post/account "doesn't violate [their] community standards."

This isn't necessarily the issue now that Facebook has linked their hardware platform (Occulus) to your Facebook account.

If you aren't buying ads on FB, you aren't a customer,no matter how you look at it.

Facebook is not an essential service; it is not a monopoly; and access to it is not a human right.

By freedom of association, Facebook ought to be able to drop users arbitrarily as they see fit.

People can't jump ship because of network effects: https://en.wikipedia.org/wiki/Network_effect That's why you see mass exoduses, like what happened with WhatsApp; there's no interoperability between Facebook and the rest of the social media space, so you lose contact with nearly every Facebook user when you leave Facebook. You can put up with a lot when it's the only way to talk to your family.

Depending on your social circles, it de facto is, and refusing to use it or being refused from it will result in being excluded from real-life social events.

The solution is to decentralize and connect our own facebooks.

2. While I don't disregard Barry Goldwater's argument against the Civil Rights Act [1], the fact is that right now we do actively interfere in the marketplace to mandate transactions not be constrained by "protected categories": race/gender/creed/etc. I don't think it's at all unreasonable to proactively extend this protection to include political alignments or even epistemic opinions.

(One of the things I find most curious about CRA, is that while most of its categories are "immutable characteristics", one's faith is more a matter of choice than genetics. If I claimed my religion was "I worship $POLITICAL_PARTY", would it then be protected? What is the scope of what defines a "faith"? In a neighboring FB banning instance [2], Lincoln Cannon even describes how his inciting post advocating decentralization derives from his Mormon beliefs.)

[0] https://en.wikipedia.org/wiki/Natural_monopoly

[1] https://www.nytimes.com/1964/06/19/archives/text-of-goldwate...

[2] https://lincoln.metacannon.net/2021/01/facebook-disabled-my-...

This is less far-fetched than you might imagine. TikTok moderators were ordered to prevent disabled users from appearing in feeds [1], which in practice is blatant discrimination against a protected class - but you'd never know about it directly since even if you figured out you were hidden from feeds, why would they answer unless they were forced to?

[1] https://www.theverge.com/2019/12/2/20991843/tiktok-bytedance...

It's pretty clear the GP wants to change the laws to make such actions by Facebook illegal. Facebook is private of course, they still have to follow the law. Are you saying you don't know specifically what the change in the laws would be (of course we can't expect the GP to have all the details worked out) or do you not understand the concept that we as a society can change the law to regulate Facebook's actions?

Many local governments, for example, use Facebook and no other mechanism nowadays.

I don't agree with this but that is reality. Until you change reality, you have to deal with it.

Now, this is bad for many reasons, but the point is in many parts of the developed world, people do think of Facebook as the internet.

Saying that a company that can literally affect global elections based on whims of political people at the top... isn't important?

the scope of Facebook in no way shape or form compares to Fortnite... unless fortnight is enforcing rules for one group but not another and fortnite affects elections and the global conversation.

This is a key failure imho. You find a long lost cousin, classmate, friend. First thing I would do is say "hey, what's your number" or "let's jump on a skype call" (come on, almost everyone has skype). Basically establish an 'independent' communication channel.

GDPR gives you the right to have any algorithmic decision explained, people just need to start using it and complaining to the commissioners.