Facebook and Google are considering banning micro targeting for political ads | Hacker News Reader

Facebook and Google are considering banning micro targeting for political ads

theguardian.com

20 pointsjonplackett18 comments

This seems like a positive move. If you can’t isolate your messages you have to be more generally appealing, and probably honest.

It also might possibly bring Facebook + Google more revenue because the groups paying for these ads have to target a blanket of people in order to hit that micro demographic they're targeting now.

+1. Also it removes the ability to send contradictory messages to different audiences.

It also removes the ability to target problematic messages to audiences which are more likely to receive those messages favorably. Which is good!

For example, if a candidate were to promise to "drive <group> out of the country", they should not be able to limit that message's visibility to audiences which exclude that group.

exactly. you force them to just deliver one message to everyone and it has to be consistent. it would force all parties back towards the political centre ground.

i hope this happens, and quick. we have an election happening here in the UK and facebook just said they're happy for the tories to use misleadingly edited clips in their ads.

a massive problem deserves a massive response and this should be it.

I've wondered if you could fix some of the problems with the modern internet / political landscape by making targeted advertising illegal.

You would still have advertising, but less incentive to collect data on individuals.

I would love to hear feedback / criticism on this idea.

> making targeted advertising illegal

how do you define targeted though? General definition would make almost all of the political ads illegal.

I think the resolution matters, as this article mentions.

Putting an ad on a bench of a district where 70% of the >100K residents belong to a certain interest group is targeted, but in a way that is obvious to anyone and fair.

Sending online or mailed ads to people who belong to the intersection of three or more groups, as identified by signals that are difficult to observe (shopping habits, search history, social media interactions) is a line I think most consider wrong to cross.

You could prohibit targeted advertising based on membership in a "protected class"[1] (race, religion, sex, etc.) - I think one of the problems with targeted political advertising on social media is that a lot of it seems created for the purpose sowing conflict between these groups. I struggle to come up with a legitimate reason a candidate would ever need to tailor their messaging based on those groups.

[1] https://en.wikipedia.org/wiki/Protected_group

The targeting can easily be neighbourhood, so you can say neighbourhood 1 (that likes Pizza) gets messaging that says candidate A loves pizza and everyone should get a free one, and also tells that neighbourhood that candidate B hates pizza. Then in another Neighbourhood that hates pizza they advertise the opposite.

You can get a reasonable approximation of that targeting in real life as well, but it is harder to hide that that is what you’re doing. Whereas with online targeting you are never aware that the other ads exist.

To me, this seems like a very balanced approach. Interested in the HN take, what level of targeting would be appropriate to allow? I can think of arguments both for and against even relatively broad targeting based on location and demographics.

Zero targeting. Why would the message change based on the audience? I can't see any legitimate use case.

i would say age, gender, location, and leave it at that!

maybe not gender actually, now I think of it, and maybe not age or location either.

force them to just broadcast the same message to everyone, and that's it.

Yeah I've been going back and forth on this. I think some location targeting is important though. For example, if you're running for mayor I don't think you should have to advertise to the whole state or country. I think the real question is should you be able to sub-target the voting populace? (i.e. I'm running for a statewide race but only want to advertise to certain cities)

If you’re up for election in region X, then your ads should have to be shown to everyone in region X.

Why age and gender?

Those are super easily (and frequently) exploited.

I initially wrote that because it seemed like the kind of thing that has historically be targetable, before online advertising - eg. by buying space in a particular magazine more prevalently read by that age group or gender. BUT I immediately took it back with another comment. I think the difference is that while you could target those specific groups back then, you still had to take out an ad that _could_ be seen by everyone else.

Reply on news.ycombinator.com