It's fairly easy to draw a line -- if at any point in the pipeline you are basing advertising decisions on data that can be linked to any one particular individual, that is targeted advertising.

If it can be shown "Bob was shown this pasta ad because he as an individual Googled 'spaghetti recipes'", that is targeted advertising.

Now, I don't agree with the article either, because I don't really see the intrinsic evil of targeted ads, but it's not hard to draw a clear line.

Advertising is showing a given set of eyes a given ad. Then how else this particular person can be shown the ad, other than by using some kind of knowledge about them?

So then the line can be anywhere between completely random and a single person. What about a group of people, like maybe two (a couple), several (a family, close friends) or a bunch (small company)?

My point is that i'd imagine many people will also have a problem with those group selections too. It's not very clear where to draw the line.

But in general, do we have a problem with targeting, or just with keeping data about individual people? What if you could target without keeping data about individuals:

- some people (e.g. Brave browser) are working on keeping the data only locally and having the browser choose the ads to be shown.

- in the future, you could have your local personal AI keeping the data and just selecting remote ads.

All laws are problematic like this. That is why we have courts, judges, and juries to sort of specific applications.

However, for the purpose of establishing a rule, this really is straight forward if we don't mince words.

The author of the article pretty clearly wants to eliminate the incentive for invasive data collection so the ad forms to be banned are the ones that require invasive data collection in order to be of value.

The precise selectors that enable Facebook's ad platform to target ads at specific types of people and provide feedback on who saw what ad are the ones to be banned. Ads placed on Facebook group pages and which do not track which specific users saw which ads, would not be banned because those are targeting a digital location rather than people who have specific characteristics.

Similarly, there is no reason to say of a given billboard "that is regionally targeted, let's ban that too" if the billboard is not currently capable of conducting eye gaze tracking on every passerby and matching car occupants to their license plate numbers and building a profile on where those people travel every day for the purpose of showing similar ads along their routes. Any billboard that could do such things, should also be banned.

The Brave browser case would probably be borderline but so long as it is not providing feedback to the ads in the form of "this person-id saw your ad-id because they previously looked at x, y, and z" then it's probably fine. If they start sharing the data they collect "locally" then that, too, should be banned. However, if they were to prompt the users with something like "would you like to submit feedback on this ad to the advertiser", that would of course be acceptable.

Brave won't share your data, we use a ZKP protocol, ANONIZE2 (2nd generation https://anonize.org/) and we're aiming to decentralize via ZK-STArKs in state channels, or better, on future Ethereum. Sharing data even with consent leaks too much, and scale players can arbitrage and abuse users across media and devices, even with Brave protecting by default. So we won't share data even optionally. Users may share by forming first-party relationships with brands, but only if they choose to do so.

We've tied the knot between user agent and user interests by (a) blocking by default, any private ads are opt-in per user; (b) all user identifying data stays on device and never goes to our servers in linkable form; (c) user gets at least as much revenue share as we get: 70% to our 30% for private tab ads; 15/15 split between us and user with 70% to the publisher if we get to filling publisher ad slots (note we use transparent, macro-auditable-on-blockchain fees, unlike most of the ad-tech parasites).

Neat! I didn't know you found a way to revenue share with users. "Make money off of your own browsing", that's a really cool hook.

Have you all considered an option to have Brave mine coins and split those proceeds? That could be an even better form of the experiment Salon.com is running.

https://www.tweaktown.com/news/60862/salon-offers-users-cryp...

The difference is in "knowing" vs "guessing".

If I am shown an ad in front of a YouTube video about how awesome guns are, it can be guessed that I like guns, and so would make sense for YouTube to show an NRA advertisement in front of that video. The "targeting" is based on assumptions about the audience, not specific knowledge. The only specific knowledge is based around the content, not the audience, and the content creator (whomever that may be) has opted in to ads.

An actual "targeted ad" counter example to this example would be if I made a post on Facebook saying "Damn, I really love guns!", and Facebook then showed me an NRA ad.

One is based on actual data about me, the other is on the content/environment and assumptions. I think that is a fairly clear difference.

You target the content and location and not the user. That is how advertising had worked throughout history,for example with TV,news papers and magazines.

If you've never heard of direct marketing, you're in for a big surprise.

I'm not sure where you stumbled into this idea, but advertising hasn't worked this way since well before the golden age of positioning...which predates the Mad Men era.

Yes absolutely it is. Cartoon channels get cereal ads. Drama shows get tax ads. They didn't know I was watching.

I'm not going to argue with you, just go check Cartoon Network's media kit.