| > People sure like to write emotionally charged posts arguing for privacy, but they're always suspiciously low on details on what bad things (actually) happened. Two bad things (random selection, because the comments below already make some really good points): 1. targeted behavioural advertising is proven to increase polarisation, literally turning people against each other. A single instance of violating someone's privacy doesn't matter as much as your single vote won't shift the result of elections. But a single vote does matter, because is a part of a bigger whole. 2. My family member suffers from PTSD acquired because of living in an abusive relationship for 2 decades. That person started a new life, but ads targeted at her and her partner more than once triggered actual panic attacks. I know this might sound ridiculous without the context. This is because that person didn't understand how clever the tech behind targeting was and assumed that the ads were related to their partner cheating on them. It's irrational, I know, but we're talking about someone who is psychologically vulnerable. I'd still say that 1. is a more important argument here, 2. just follows the line of thinking presented in your comment. (the main problem behind 2. is that person's mental state and the actions of their abuser, yet the amount of suffering that could've been removed is not negligible.) > Even in this case with phone numbers and other data leaked, so what? What harm do data leaks cause? Cambridge Analytica, voter manipulation, bias in behavioural targeting, increased polarisation in media--please Google these queries and educate yourself. There's a tonne of resources on the subject, including peer reviewed academic papers. |
Can you provide some evidence for this please? Certainly, filter bubbles make it easier for people to radicalise themselves, but I've not seen very much evidence that it's specifically the advertising.
And polarisation in (US) media has been underway since long before Mark Zuckerberg left elementary school.