3 ms·

Even in systems with automated processes, the bare minimum for communication should be a log of activities that is visible to all parties (e.g. account owner an

Even in systems with automated processes, the bare minimum for communication should be a log of activities that is visible to all parties (e.g. account owner and company). The log should remain visible for quite awhile even after a severe action such as a ban. And in such a log, every action taken by an employee should be marked with an employee number, even if it is a number that only means something inside the company; and every action taken by an algorithm should be marked with an identifier for that program. It would be a lot easier for everyone involved to understand what the hell is going on if you could see entries like: - <YYYY-MM-DD> algorithm XYZ flagged account due to suspicious activity [link] - <YYYY-MM-DD> employee 123456 verified accuracy of terms violation [link] - <YYYY-MM-DD> algorithm ABC notified account owner via E-mail of violation [link] - <YYYY-MM-DD> algorithm ABC auto-banned account Instead, it seems for a lot of tech companies the communication amounts to ONE E-mail basically listing all the lazy non-effort they went to before ruining everything for you with no recourse.

mattlondon 8y ago
With respect, I am 100% sure that Google does this (no insider knowledge here - just conjecture on my part). I am completely sure that Google - a company that lives and dies by data - keeps logs. But there is zero value sharing this with "customers". If you shared this and it said "Algorithm/Employee #12345678 linked this account with previously banned account" where does that get you? It doesn't get you any further - all it gets you is another thing to argue about, but they've already made their decision and you agreed to their TOS about how they can cancel your account at any time, so you've gained nothing really: Google still don't want to do business with you, sorry.
crazygringo 8y ago
The piece of the puzzle that you're missing is that people use accounts for fraud, and if they find out why an account got banned, they'll know which behavior triggered it and will learn to avoid that behavior in the rest of their accounts and future accounts. It sucks, but automating sharing reasons for flagged actions will be self-defeating. Ultimately it needs to be reviewed by a person at the company who is equipped to judge the situation with more intelligence and discretion to determine if it's really fraud or not.
- megous 8y ago
  > The piece of the puzzle that you're missing is that people use accounts for fraud, and if they find out why an account got banned, they'll know which behavior triggered it and will learn to avoid that behavior in the rest of their accounts and future accounts. Compliance is the goal, no? How's it self-defeating if you ban non-compliant behavior and only way around it is to basically get more compliant with the rules? How will you get compliance if you don't share details with clueless violators, so they can avoid the mistakes in the future? That seems self-defeating, because this gets rid of clueless, but potentially valuable developers.
- crankylinuxuser 8y ago
  > The piece of the puzzle that you're missing is that people commit crime, and if they find out why they got in trouble with the law, they'll know which behavior triggered it and will learn to avoid that behavior. > It sucks, but automating sharing reasons for illegal actions will be self-defeating. Yes I changed your text from corporate "illegal" to government "illegal". We wouldn't let THAT fly in a court of law. We have rules on that, like the 4th and 5th amendment. Right to have the charges read before them; right to face their accuser; right to a fair trial; right to be secure in their houses/papers/effects So what are the charges? "We will not tell you what you did." Who's the accuser who said I did the thing? "It cannot be in the courtroom. It is an algorithm. Secure in your effects? You waived them (unilaterally) in dealing with a monopoly. Too bad, soo sad. Fair trial? SURELY YOU JEST!

mattlondon 8y ago

With respect, I am 100% sure that Google does this (no insider knowledge here - just conjecture on my part). I am completely sure that Google - a company that lives and dies by data - keeps logs. But there is zero value sharing this with "customers". If you shared this and it said "Algorithm/Employee #12345678 linked this account with previously banned account" where does that get you? It doesn't get you any further - all it gets you is another thing to argue about, but they've already made their decision and you agreed to their TOS about how they can cancel your account at any time, so you've gained nothing really: Google still don't want to do business with you, sorry.

crazygringo 8y ago

The piece of the puzzle that you're missing is that people use accounts for fraud, and if they find out why an account got banned, they'll know which behavior triggered it and will learn to avoid that behavior in the rest of their accounts and future accounts. It sucks, but automating sharing reasons for flagged actions will be self-defeating. Ultimately it needs to be reviewed by a person at the company who is equipped to judge the situation with more intelligence and discretion to determine if it's really fraud or not.

megous 8y ago
> The piece of the puzzle that you're missing is that people use accounts for fraud, and if they find out why an account got banned, they'll know which behavior triggered it and will learn to avoid that behavior in the rest of their accounts and future accounts. Compliance is the goal, no? How's it self-defeating if you ban non-compliant behavior and only way around it is to basically get more compliant with the rules? How will you get compliance if you don't share details with clueless violators, so they can avoid the mistakes in the future? That seems self-defeating, because this gets rid of clueless, but potentially valuable developers.
crankylinuxuser 8y ago
> The piece of the puzzle that you're missing is that people commit crime, and if they find out why they got in trouble with the law, they'll know which behavior triggered it and will learn to avoid that behavior. > It sucks, but automating sharing reasons for illegal actions will be self-defeating. Yes I changed your text from corporate "illegal" to government "illegal". We wouldn't let THAT fly in a court of law. We have rules on that, like the 4th and 5th amendment. Right to have the charges read before them; right to face their accuser; right to a fair trial; right to be secure in their houses/papers/effects So what are the charges? "We will not tell you what you did." Who's the accuser who said I did the thing? "It cannot be in the courtroom. It is an algorithm. Secure in your effects? You waived them (unilaterally) in dealing with a monopoly. Too bad, soo sad. Fair trial? SURELY YOU JEST!