7 ms·

> Based on the frequency with which this sort of post gets significant traction it's pretty clear that no one at Google gives a crap about this issue. Lots of

> Based on the frequency with which this sort of post gets significant traction it's pretty clear that no one at Google gives a crap about this issue. Lots of other reasonable explanations that match the available evidence and involve Google caring about this issue. For example, their hands could be tied by legal forces beyond their control. Or they could care a great deal, but perhaps they have been unable to resolve this issue with an acceptable false-negative level on actual bad actors. Another possibility is that the ban was well-earned and the article's author knows this and simply has not disclosed the fact to the readership.

noneeeed 5y ago
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.
- aaomidi 5y ago
  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.
- asdfasgasdgasdg 5y ago
  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.
  - heavyset_go 5y ago
    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.
  - Beltalowda 5y ago
    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".
    - asdfasgasdgasdg 5y ago
      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.
      - Beltalowda 5y ago
        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.
- VPenkov 5y ago
  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.
vlunkr 5y ago
But you see these stories again and again about Google specifically. The creator of Terraria was banned a while back and was also unable to get it corrected or get a real human to respond. You'd think if they are terminating accounts that are bringing in money they could put down the automation for one second and actually do the right thing.
- dielll 5y ago
  Maybe google only makes money from ads, other projects are hobby projects to them. hence why Admob has great customer service
tablespoon 5y ago
> Lots of other reasonable explanations that match the available evidence and involve Google caring about this issue... Or, more likely: they "care," but not enough to actually spend the money to fix the problem (e.g. quit relying so much on algorithms, and staff a department of real people to apply some human judgement). My understanding is there's a cultural problem at Google, where they think stuff like this must be done algorithmically, even if it's not a actually a job for an algorithm.

noneeeed 5y ago

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.

aaomidi 5y ago
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.
asdfasgasdgasdg 5y ago
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.
- heavyset_go 5y ago
  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.
- Beltalowda 5y ago
  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".
  - asdfasgasdgasdg 5y ago
    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.
    - Beltalowda 5y ago
      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.
VPenkov 5y ago
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.

vlunkr 5y ago

But you see these stories again and again about Google specifically. The creator of Terraria was banned a while back and was also unable to get it corrected or get a real human to respond. You'd think if they are terminating accounts that are bringing in money they could put down the automation for one second and actually do the right thing.

dielll 5y ago
Maybe google only makes money from ads, other projects are hobby projects to them. hence why Admob has great customer service

tablespoon 5y ago

> Lots of other reasonable explanations that match the available evidence and involve Google caring about this issue... Or, more likely: they "care," but not enough to actually spend the money to fix the problem (e.g. quit relying so much on algorithms, and staff a department of real people to apply some human judgement). My understanding is there's a cultural problem at Google, where they think stuff like this must be done algorithmically, even if it's not a actually a job for an algorithm.