Hacker News Reader

Comment by noneeeed | Hacker News Reader

What sort of legal issues would prevent Google actually providing some information on the reason for the ban, of prevent it actually engaging in any kind of useful communication?

My issue with these bans is not the bans themselves, but the Kafkaesc process in which you are supposed to argue why you think their decision is wrong without knowing what they based the decision on.

View on HN

I'm not in a position to comment about what sorts of legal issues there might be. But Google is hardly the only company that hands down bans with scant explanation. It's endemic. I've actually not heard of a large company that gives detailed explanations in these sorts of situations. That leads me to believe there is a good reason for avoiding such explanations, since I would expect some variation in company behavior otherwise.

The reason is that it's expensive and companies don't want to waste resources verifying whether or not their bans followed their policy accurately or not.

For every account banned, one million new users will enter the Google ecosystem. Bans cost them nothing in the long run, but having reviews to justify bans would.

I suspect it's because a lot of these automated bans are done without any clear actual reason beyond "our machine learning model found something, but we can't really get a coherent picture of why".

Last week I created a brand new Facebook account as I recently moved to a new city and am somewhat desperate to meet people, and Facebook events can be useful for this (as much as I loathe Facebook, my mental health is a bit more important right now). I got banned after about 2 minutes. I didn't actually do anything on the platform. I managed to click to about 2 or 3 pages.

What set this off? Who knows. Maybe some browser settings? Previous user(s) of my IP address? Those pages I clicked on? Something else? It surely wasn't anything I actually did as I didn't really interact with the platform at all.

I suspect that if I asked some high-up Facebook engineer to look in to it they wouldn't be able to actually give me an answer either beyond "The Algorithm determined there were risk factors".

I doubt it, personally. At least in the case of the article. I have no special knowledge of this, but tenure would be one of the first override parameters I would use to trigger human review. It seems unlikely to allow an account that has been active and making money for six years to be banned without a human analyst in the loop.

As to your FB ban, I'm sorry that happened but I have no insight to share.

I'm not so sure about that; like you said in your previous comment: the scale of all of this is huge, and the amount of money these companies are making is similarly huge. It's genuinely a hard problem, and they can afford to lose this business.

Was that a complete ban, or was it an attempt to get your phone number and other personal details?

Instagram has always banned my accounts for suspicious behaviour on whatever page is first used after the account is created - the only way they offer to progress is to give them a phone number. It's a pretty transparent grab for more info to sell.

Twitter does this, as well.

They did ask for my phone, but I think that's okay; it's somewhat effective at preventing spam/abuse.

Funny enough, when I checked after posting my previous comment I could log in and it was all good. I didn't do anything, and now I got an email saying "Your Facebook account has been disabled. This is because your account, or activity on it, doesn't follow our Community Standards". All I did was log in, view the home page, and close the tab.

I requested another review :-/

I think it's not strictly legal issues but one of two things.

The first is what I've heard being called "a healthy dose of obscurity". With a large enough pool of users, there is a finite number of behaviors that would trigger a certain ban rule. To avoid having their triggers reverse-engineered, they say as little as possible.

The second is is reasons being subject to interpretation. The risk there comes in a variety of flavors from PR risks to lawsuit risks when a developer are banned for X but there's an important nuance to X that the developer ignores by choice or otherwise.

I don't mean to defend anyone here, I'm just hoping that someone smarter than me will tackle this somehow.

If they receive a court order it generally comes with a gag order too.

Theres nothing stopping them from coming clean after the period of that order has dropped though.

Reply on news.ycombinator.com