Hacker News Reader

Presenting both options on equal terms is hard to measure, Apple is calling it "Tracking" when other (Facebook's) advertising networks do it, but they are calling it "Personalized Ads" when their advertising network does it.

Apple doesn't mind if users turn off the data feed to Facebook, but obviously they don't want users disabling the same data feed to their advertising network.

It’s very possible to personalize ads without personal information.

It’s not possible to de-personalize Facebook, its whole knowledge graph is built around identities. Permission to use that data can be managed internally but it’s always gathered and processed initially as personalized.

Ok, but is that actually what Apple is doing? If the distinction is purely hypothetical, it's not relevant.

It's OK to be angry that Apple exempts themselves from privacy requirements, but that doesn't mean that users aren't showing their true preferences when they opt out of tracking. The options in regards to Facebook are still being presented on equal terms.

It would of course be good for Apple to call their own behavior "tracking" as well, because that's what it is; that is the most accurate word for what is happening. "Personalized ads" do not accurately reflect what is happening because ads can be personalized in lots of different ways, only some of which involve user tracking.

But I don't think that has anything to do with Facebook's argument. Facebook's argument here isn't "everyone wants to exploit users, but only Apple is getting to do it", their argument is that users like tracking, that the tracking Facebook does is not exploitative. Well if that's the case, then if you show users a dialog that asks them whether or not they want to be tracked and there's a "yes" button and a "no" button, then users who do want to be tracked will click "yes."

When I say that the options are shown on equal terms, I mean that neither option is privileged for Facebook. It's exactly the same amount of work for a user to opt into Facebook tracking as it is for them to opt out if it. That means that Facebook can't blame application defaults for consumer behavior, consumers are actually making a real choice here.

Calling it tracking is more honest. Maybe personalized ads is a derivative of tracking data. It since it requires tracking to produce personalized ads, asking if you want to be tracked is the honest approach. Apple is trying to avoid dark patterns that Facebook was essentially built upon.

Plus I have zero confidence that if I were to turn off what Facebook calls personalized ads, they would stop tracking me.

That’s because Apple’s ad personalization doesn’t revolve around cross-site tracking. It’s based on demographic information you provide them and preferences you select within Apple’s universe of services. The only place it arguably starts to cross into the realm of tracking your activity is when it determines your genre preferences based on which movies/music you buy on iTunes or which publications you subscribe to in Apple News. But I don’t think anyone would be surprised by or offended about Apple Music recommending you more music that’s akin to other music you like.

The distinction Apple makes is that iAds don't "track" you because they don't use cross-site or cross-app tracking in third party services the same way Facebook or Google do.

Apple is of course still performing data collection for targeted iAds, but that's the reason they don't call it "tracking."

If you go to Settings > Privacy > Apple Advertising on iOS it tells you so:

"The Apple advertising platform does not track you. It is designed to protect your privacy and does not follow you across apps and websites owned by other companies. You have control over how Apple uses your information."

Note that I'm not necessarily agreeing that what Apple does doesn't fall under the description of "tracking." Just pointing out why Apple calls what Facebook does "tracking" but not their own iAds.

You assume that Apple is telling the truth. The cynic in me thinks that they narrowed the definition of tracking and follow in a legal sense but not in the way we would think of it. Or they are outright lying. It is also possible that it isn't apple doing it but a third party doing it on Apples behalf.

It's possible they're lying, that's true, but I think it's unlikely a secret like that could be kept long term. Employees have no reason to remain loyal when no longer employed by Apple and partner company employees would have to know about it too. The reputation damage would be catastrophic, so IMHO it seems unlikely they would take such a reckless risk.

So it's a matter of making a call between using vendors we know for a fact are tracking you, some of which have lied about it (see, hard to keep these things secret), or one that at least claims not to and has a lot to lose if caught out.

The burden of proof would be on you to prove Apple is using third parties to track its users in that case. Without proof it's just a goofy conspiracy theory.

My first question would be... why? What's their motivation?

Facebook has to track users to make money, it's their entire business model.

The same is not true of Apple, who make almost all their profits from hardware sales and are worth over $2 trillion.

Going out of their way to get a third party to secretly spy on users on their behalf sounds like far more effort and risk than it's worth to me. It's not like they're desperate for extra cash.

And it's not like it'd be impossible to find out if it was indeed the case. iOS gets reverse engineered all the time. I can see all the traffic my iPhone sends through my DNS logs. If this hypothetical third party was embedded in iOS spying on user activity, it'd be spotted quite easily and quickly.

So I ask again, why would Apple do that? Aside from being a total PR nightmare it'd be straight up illegal.

I trust Apple because I don't see what reason they have to lie when they've built a multi-trillion dollar business without tracking their users already.

> Or they are outright lying.

It seems unlikely Apple is going to outright lie about something like that considering it could open them up to serious legal liability (CCPA, GDPR, etc).

I think the problem is that facebook never had to fight for tracking. Everyone just just got tracked and fb could do what they want. They never cared whether it was really useful, they just tried to show as much relevant ads as possible. A complete disregard for the user-experience thanks to it's monopoly, which also shows in fb getting unpopular (at least in my circles).

I don't know. When I first heard people talking about dark patterns in UI design, Facebook was the platform that immediately sprung to my mind.

It was always laughable (?) the hoops you had to go through to turn on various aspects of privacy in their settings. And it was constantly changing, requiring users to constantly go back into settings to turn off yet another flag that was introduced default-on.

They seem to have known very early on that user's would not appreciate their crap.

I'm a bit of an ideologue on this point, but I would argue it's not just Facebook -- nearly all modern advertising is divorced from the idea that ads need to provide value to consumers. At best, ads are treated as an annoyance that users put up with in order to access content that they want.

But the advertising industry desperately doesn't want to acknowledge that. They want to stick their heads in the sand and ignore that most users don't want to see anything that they're making. They want to argue that most users are benefiting from ads, even though that's clearly not true.

A system that aligned incentives between advertisers and consumers, where advertisements themselves actually needed to justify themselves and inform/entertain users enough that they provided positive value to the viewer -- that system would look very, very different from the advertising world we have today.

I think that disconnect is why you see people in the advertising industry fighting about ideas like users' Right to Filter, or why they're so paranoid about users having more control over which ads they see. Advertisers know that nobody likes ads and that for the average consumer, their ads aren't providing any kind of actual value. There's no incentive for most online advertisers to make ads that improve the quality of anyone's life -- the biggest incentive in the current system is to alter consumer behavior against their will regardless of whether that ad helps or hurts the viewer.

It’s an unwarranted assumption that people given a question to answer always understand what it’s asking and answer truthfully. Another possibility is that many users won’t understand or care much and will just click at random to make the dialog go away. Or they might use default rules like saying “no” to anything they don’t understand.

Since we can’t ask users what they meant, it’s not something we can know.

For more skepticism about the over-interpretation of survey results: https://carcinisation.com/2020/12/11/survey-chicken/

So this argument reduces to: “People might not understand a question which is important that they understand. So let’s just not ask them in case they misunderstand and assume they answer in a way that is beneficial to us.”

Does that sound right?

Not at all. I didn’t suggest any alternative. Given the controversy, I’m not sure there’s a better one.

But you might compare to what browsers do these days for visiting websites with fishy-looking certificates. They don’t just put up a dialog, they heavily discourage it, because the assumption is that many users would say “yes” just to get rid of the dialog.

And with the vastly-majority of sites now using HTTPS certificates, the "invalid SSL certificate" page has become exceptional enough to have me actually read the "why" is the browser complaining so much, instead of just clicking on the "eh, whatever, I know the risk, drive me in!" button, which I would have clicked without remorse 15-20 years ago.

In general this is a an argument that users shouldn't be given a choice because "you" know better than some/many users. So "who" knows enough to force an option on everyone because some users might not understand this choice between having Facebook track information about them or not?

Would you argue that increasingly targeted advertisement is enough of a net good that the choice to avoid it should be gone?

I think Apple’s approach is reasonable and I don’t have a better way. I’m just warning against over-interpreting the results as a strong preference.

That could be a good argument for Facebook explaining why they ask for a permission before prompting, as many apps do. but it’s hard to believe this is their reasoning when they’ve strenuously avoided both asking permission and honestly explaining what they intend to use it for. They’ve had years to be more transparent and have presumably thought about why they’ve spent millions fighting it.

That’s also an argument against allowing everyone to vote.

It’s a fully general argument against trusting knowledge gained from surveys, and I do think we should be skeptical that voters always know what they’re doing. Particularly for elections they aren’t paying attention to.

The “will of the people” is something of a myth. Votes go in and the algorithm makes a decision. Do you feel represented when the decision isn’t what you wanted?

But it’s not an argument against elections because I’m not claiming I have a better idea. We do need some peaceful way of making decisions that people will accept. A legitimate, accepted process beats chaos, and new processes aren’t easily adopted.

Elections are less about finding the optimal will of the people and more about creating an adversarial environment for politicians in order to prevent capture by a single party.

Facebook has regularly taken the position that users know what they're opting into when they use Facebook.

If they want to take the position now that their users don't understand what "tracking" even is, then Facebook was lying before when they said they had informed consent from their users to do that tracking.

The advertising industry regularly wants to have its own cake and eat it too on this point. They want to say that users understand and consent to the system, but also that users are too uneducated or lazy to acknowledge that consent in any way.

And to a small extent, there is a grain of truth here: most users don't understand the full extent of what tracking is and how companies are targeting them. But that's not a point in the advertising industry's favor. Companies like Facebook acknowledging that users don't know what they're opting into or out of only makes it more obvious just how inadequate their click-through TOSes and privacy FAQs always were.

>So Facebook wants to argue that it's a worse experience for consumers if they're not being tracked, but the truth is that they know a large portion of users don't want the experience that Facebook is offering.

I think this is true for a lot of people on social media. They're there for the people and topics they care enough about to explicitly opt into. Ads and recommendations can fuck right off.

This describes it pretty much spot on. It’s the change to make it transparent and clear to end users. One of the core accusations against google and Facebook is that they have been doing this on the sly and in-transparently, as is well documented in the book surveillance capitalism by zuboff.

They have done this for decades, with google paving the way, by slowly and surely chipping away at peoples privacy. People then just give in over time, amplified by single sided terms of service that allow them to pretty much change anything at any time.

This is hopefully one of the major turning points where this kind of behaviour will no longer be tolerated. Expose privacy. Make it explicit. Let people actually decide.