3 ms·

This is from 2012 and uses much less sophisticated tracking than what happens on the internet: https://techland.time.com/2012/02/17/how-target-knew-a-high-schoo

This is from 2012 and uses much less sophisticated tracking than what happens on the internet: https://techland.time.com/2012/02/17/how-target-knew-a-high-school-girl-was-pregnant-before-her-parents/ https://techland.time.com/2012/02/17/how-target-knew-a-high-... The police don't have to do the work of correlating the data themselves. Would you stake your freedom on the idea that there's not a data broker somewhere buying up data and correlating it to create a "pregnancy score" for advertisers that might also be useful for law enforcement?

legitster 4y ago
Okay, but there's a difference between Target throwing darts and accidentally hitting a target[sic] and legal grounds to press charges. I administer marketing systems for a living, and on a good day we hit 1 target out of a hundred. People really overestimate how sophisticated the data is. Police could have gone around buying phone books to look for parole jumpers - but the cost/benefit ratio on following up on every potential name would have been absurd. This would be a whole other magnitude above that.
- AlexandrB 4y ago
  Perhaps I'm being cynical, but all it takes is an "expert witness" from the analytics firm in court saying that the chances the defendant was not pregnant are less than "1 in 10,000" or something. How good the technology actually is might not even matter to a jury considering the hype around ML and AI. > Police could have gone around buying phone books to look for parole jumpers - but the cost/benefit ratio would have been absurd. This would be a whole other magnitude above that. I'm not sure the cost will matter much when we're talking about an ideological crusade like this.
  - legitster 4y ago
    I think the distinction here is that this data could be used as evidence if they already have a suspect. In that case it's much easier to get you to turn over your phone and usernames and activity and testimony, etc. But I think the risk of people being "caught" with this data is absolutely minimal. > I'm honestly not sure the cost will matter much when we're talking about an ideological crusade like this. Okay, but even in like the darkest days of prohibition when the federal government had a literal mandate to round up every person in the US for illegal drinking, not even a fraction of a percent of the obviously guilty population was even charged with a crime. At a certain point, laws themselves are a legal fiction if they are unenforceable and police will only ever focus on a few high profile cases. So I guess I am cynical myself, but also trying to be rational here.

legitster 4y ago

Okay, but there's a difference between Target throwing darts and accidentally hitting a target[sic] and legal grounds to press charges. I administer marketing systems for a living, and on a good day we hit 1 target out of a hundred. People really overestimate how sophisticated the data is. Police could have gone around buying phone books to look for parole jumpers - but the cost/benefit ratio on following up on every potential name would have been absurd. This would be a whole other magnitude above that.

AlexandrB 4y ago
Perhaps I'm being cynical, but all it takes is an "expert witness" from the analytics firm in court saying that the chances the defendant was not pregnant are less than "1 in 10,000" or something. How good the technology actually is might not even matter to a jury considering the hype around ML and AI. > Police could have gone around buying phone books to look for parole jumpers - but the cost/benefit ratio would have been absurd. This would be a whole other magnitude above that. I'm not sure the cost will matter much when we're talking about an ideological crusade like this.
- legitster 4y ago
  I think the distinction here is that this data could be used as evidence if they already have a suspect. In that case it's much easier to get you to turn over your phone and usernames and activity and testimony, etc. But I think the risk of people being "caught" with this data is absolutely minimal. > I'm honestly not sure the cost will matter much when we're talking about an ideological crusade like this. Okay, but even in like the darkest days of prohibition when the federal government had a literal mandate to round up every person in the US for illegal drinking, not even a fraction of a percent of the obviously guilty population was even charged with a crime. At a certain point, laws themselves are a legal fiction if they are unenforceable and police will only ever focus on a few high profile cases. So I guess I am cynical myself, but also trying to be rational here.