Hacker News Reader

Comment by cm2012 | Hacker News Reader

cm2012

Okay, do you have an example of someone being hurt by tracking and data collection done by Facebook and Google?

Sure. FB knows your ethnicity. In violation of US federal law, this information is used to show you different apartment ads. This makes it harder for black people to live in integrated communities.

Similar things happen with jobs and ethnicity, jobs and gender, with jobs and age discrimination, etc.

FB does enforce the fair housing act and other laws. The platform is actually quite strict nowadays and catches lots of innocent ads in their filters for protected categories. Ad platforms should follow the law, that's working as planned.

FB may do so now but that's only because of lawsuits as recently as 2019. It wasn't something they volunteered to do.

But you asked how targeted ads hurt people, and I showed a fairly recent example. Something that wouldn't exist without targeted advertising. How does that not prove my point?

Its not like FB has ever let advertisers "exclude black people" before. The critics said it may be possible that protected categories get less ads for housing, employment, etc due to algorithmic bias. That is a subtle danger, and honestly hard to technically implement at scale, but I think FB has always given a good faith effort in this area.

If by "not ever let advertisers exclude black people" you mean "after March 2019 has not knowing let advertisers exclude black people" than FB and the NYT would agree with you: https://www.nytimes.com/2019/03/19/technology/facebook-discr...

No, you have never been able to target by race. NY times is referring to algorithmic targeting that incidentally correlated to race, which FB cleaned up in 2019.

There is no difference between "filter out black people" and "filter out X, (X correlates 99% with black people)". And I mean that legally, morally and practically.

Here's a list of ways around "cannot filter by race". It took 2 seconds on google: https://www.searchenginewatch.com/2011/12/06/how-to-target-e...

But, you seem to be shifting the goalposts pretty significantly. So, do you agree with the base point that targeted ads can produce pretty bad results?

I'm not them but from the top of my mind, an example:

1. Facebook tracks users across the internet to enrich the profiles of their users

2. Data leaks.

3. Users are now in trouble.

I can come up with many more if you ask and then give me a sec.

This doesn't actually happen though. When FB is hacked its the same banal hacks anyone else has, phone numbers, etc.

Cambridge analytica, brexit (economically) hurt a lot of people. Similarly the regime changes in 3rd world countries that these guys engaged in probably physically hurt quite a few people.

I mean there are so many examples it's difficult to believe this is a genuine question. It sort of like asking in 1984 did anyone really get hurt by big brother watching?

Cambridge is a nonsense story that people who know fb ads at all laugh at. It didn't impact the election or any election in any way.

I was shocked at the Cambridge story because I thought everybody already knew that hundreds of thousands of 3rd parties had full access to that sort of data, with the only protection being Facebook had a policy that you weren't supposed to make a copy & pass it around.

The collection is per se harmful as long as warrants exist.

Beyond that, it's an open secret (as in: it's been mentioned several times in available documents, but never deliberately disclosed or extensively discussed publicly, so far as I know) that since at least the mid or late 00s the US government has had contracts with multiple major tech companies that have a high level of access to citizen Internet traffic to basically search their databases of Internet activity at will. Which companies these are, I'm not sure—I suspect it's mostly telcos, personally—but it's another reason the existence of these datasets is inherently dangerous, and that they should not be permitted to exist at all, no matter who holds them.

The government doesn't go to fb when it wants someone's website history. It goes right to the ISPs.

What about your chat history then? Social graph? Statistical model of your behavior?

https://news.ycombinator.com/item?id=24719416

That has nothing to do with ads, Google would turn that over even if sub supported.

> do you have an example of someone being hurt by tracking and data collection

Fair enough, I should have chose wording better. In any case I would posit that the police using search engine history in evidence for cases is not a bad thing.

There are so many examples readily available that I have a hard time believing any HN user claiming to be unaware of them is anything other than willfully ignorant. It’s probably discussed here every single day if not multiple times each day.

A simple search here for “facebook data” will get you started: https://hn.algolia.com/?q=facebook+data

See my other comments on responses to specific claims. 99% of HN comments on the matter talk about irrelevant or unrelated risks.

Reply on news.ycombinator.com