Apart from seeing targeted ads, why is data collected by companies about users a bad thing?
In: Technology
Another form of misuse of the data: imagine there’s a correlation between your shoe size and your history of committing crime. The correlation is strictly non-causal (shoe size doesn’t CAUSE more crime), but the pattern is robust (there is a real correlation).
A company can sell this information to employers who can use that information to discriminate against applicants because it leads to a tiny but positive effect on quality of hires. The downside is that it’s still pretty random (a lot of people are being discriminated against despite not actually having a criminal record, only for the correlation between shoe size and criminal record).
Further, imagine they don’t know your shoe size, But the algorithm can guess your shoe size based on the third letter of your last name with a small, but noticeable degree of accuracy (above chance). So now you’re being discriminated against for your criminal record that was a guess based on your shoe size that was furthermore a guess based on something else unrelated.
There’s lots of “black box” algorithms that can be made. Black box refers to the idea that the companies know what variables go into the box (X) and it can see the effect (output, Y) of the box, but, to some degree, don’t understand ***how*** X leads to Y. And there’s plenty of bs things that can happen from this if they’re not careful or don’t care. E.g., maybe Black people are more likely to have a last name with the third letter B, and their algorithm ends up suggesting discriminatory advice that ends up mostly affecting black people.
Finally, with all this data, they can totally invade your privacy. They can guess your exact address, nationality, health history, whatever using bits and pieces of information with some accuracy because of their large datasets and modern techniques for analyzing data. And then my personal info gets sold or stolen. In a way, it’s as simple as knowing Gonzalez is a common hispanic last name. But because of large datasets, it goes way, way beyond that.
And imagine how easy it would be to steal from someone’s house if they hacked some phone service and got my GPS data. They could guess what times I’m not at home (pre or post corona-times haha).
Latest Answers