Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Can you elaborate about what you define as "used against you" even if it is entirely inaccurate? What is the use case of inaccurate data with which you are concerned?


> Can you elaborate about what you define as "used against you" even if it is entirely inaccurate?

Hypothetical: $company you're trying to use needs to "verify" you using $inaccurateData from $vendor.

You're absolutely screwed if the verification questions you're asked are relying on the "polluted" answers

Similar vein: if the "polluted" data indicates you might be gay or replublican or musilim or into some seriously unhealthy lifestyle choices like smoking and $someCompany decides that you're a smoker and therefore your too risky to insure.

no data >>> polluted data.


Serious question: Who is using browser data for verification?! It's alarming to me that this is even a hypothetical scenario. All identity verification systems I have ever used in the US have been through a credit agency or something similar. I can't imagine any use case that would use your browser history or ad data for these purposes. Do you have a real-world example?


Or, if state level actors are looking at your data they are buying from companies, the appearance of intentionally corrupted data could invite more scrutiny.


If state-level actors are looking into your data with any amount of individual scrutiny you are already fucked, this is a ridiculous reason to not use ad nauseum.


You're looking at this backwards...

Imagine being in China where they tend to watch you and make profiles on you. Then suddenly the profile of who you are goes completely random. Is it possible this gets the attention of state-level actors where you had none before?


> You're looking at this backwards...

Exactly.

Any argument that boils down to "but the profile is basically useless" will always be inferior to "there is no profile"


Poisoned data would be useful in the fight but yeah, "garbage data looks like someone else's" is certainly superior to "garbage data looks like it's yours".


I guess in the long term it depends how good the profile builders get at anomaly detection, and at which scale we're talking about.

While many states in the US have laws against it now, for awhile there companies were basing if they would hire you based on your social media profile. Having no profile at all may exclude you from getting a job. Or, when I went to get credit for my the first time in my later 30's. I had always been a cash buyer before then, and proof of my existence beyond my ID was sparse, the guy on the other end of the line was like "Did you even exist before yesterday?"


Isn’t the alternative in the verification case just that you don’t get your account because they can’t verify you?


Verification questions are based on credit history - addresses, vehicles, family members, etc - not browsing history so this is a non-issue.


See using clicking rate or browser signature to deny you service?

See Aliexpress accounts becoming unusable based on ... well, who knows what?

See Google accounts becoming unusable based on ... well, who knows what?

Not the same as identity verification but a significant problem for your account still. Often fatal. It is an issue.


Another example that I think captures the spirit of autoexec’s point is credit fraud.

Are you the one taking out credit cards and potentially tanking your credit score? No.

Does it still negatively impact your life? Yes, because the information landlords/banks receive from credit unions only shows the low credit score.

Do the banks/landlords care about the fact that it’s fraud? No.

It’s ultimately YOU who has to do all the leg work to report the fraud, make sure that your credit history is fixed, and that your credit is frozen as a deterrent to for future fraud issues.


It's in the comment they linked to. Good points, to be fair.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: