[thinkloop]

Part of op's point is that this exercise only works when you "think bigger-picture", but in practice it hasn't been panning out. Things have already been leaked, and what has the material damage been? Doesn't seem like much so far, especially weighed against the benefits.

I personally don't agree, but I believe that was op's point. I think the low probability dystopian possibilities are severe enough to outweigh the casual benefits.

[rendall]

<< I think the low probability dystopian possibilities are severe enough to outweigh the casual benefits. >>

I agree, although perhaps I am hypocritical, because I use gmail and google maps. I have jettisoned FB though.

<<Things have already been leaked, and what has the material damage been?>>

I think you're probably right, that this was their main point. To address it, I'll just briefly point to

(Edit: I may not have directly addressed their point here, but I dont think personally that I need to actually see something bad happen on US soil to assume it can't or won't happen if there is no principle to prevent it)

- the myriad oppressive regimes and violent drug cartels that use even publicly-available information to target political opponents.

- the ability of even casual users to create advertisements targeting (or excluding) ever-smaller circles of identifiable traits

- vast and secret subpoena powers by the US government and her allies

- severe political polarization centered on the US, in which the government and its powers flip poles every few years: what is okay for "our guy" to do (e.g. secret assassination lists, secret trials, rendition) will be available for "their guy" in a short while.

- AI being to identify ever-accurate and surprising correlations between lifestyle choices and political views.

Hopefully, nothing bad will happen.

[piokoch]

Information about can be used against you in many different ways: you can get higher prices in online stores (if, for instance, online store gets data that you are a rich person).

You can get higher insurance price if you insurer knows "something" about you or you can be denied insurance at all if they suspect that you are, say, an addict (and this does not have to be true, some "clever" AI will just qualify you as an 34% addict and good luck if such nonsense data, derived by some SF "move fast, break things" startup, will be interpreted in a sensible way by the insurer).

From the information Google/FB has they can derive that someone has a mistress and those who will learn that might try to blackmail such person. Or someone is looking for a new job and informed person can sell this konwledge to one's current employer.

[Karunamon]

This is non-responsive to GP's point, though. This is all hypothetical. Where's the actual, concrete, demonstrated harm?

"Someone in the future might do something untoward" is not concrete.

[rendall]

GP's point was that they could only think of bad things happening if these companies leaked personally identifiable data. If I have not successfully addressed that specific point, demonstrating that there are worse things, then I do not know how to.

[vageli]

Facebook used to have "potentially treasonous" [0] as a user attribute that advertisers, etc could use to target users. This is only the beginning. I have no doubt that more conservative regimes have used their legal systems to compel Facebook to hand over all information on all gay users, for example.

https://www.forbes.com/sites/kalevleetaru/2018/07/20/faceboo...

[sciurus]

Do you have another source? That link 404s.

[vageli]

Goodness, I am a dunce and miscopied on my phone. Forgive me!

https://www.forbes.com/sites/kalevleetaru/2018/07/20/faceboo...