Publisher opt-out ?

[edouardletort]

Hello everyone,

Publishers might not want browsing history from their content to be used by the browser to build Floc.
Can publishers opt-out from contributing to FLOC? If so, how would opt in/out be setup and what would be the limitations ?

[jkarlin]

Thanks for opening the issue. I think that this is a good point. One approach we've been considering is making FLoC opt-in by the publisher page via script or headers. Ad tech running on the 1p page is likely to opt the publisher page in, so to prevent that there would also need to be an explicit opt-out response header that script can't override.

[AramZS]

I support this concern and agree that there needs to be an explicit opt-out that can't be overwritten by downstream agents.

[dmarti]

There are sensitive vintage pages out there (such as old forums and mailing list archives) that don't receive a lot of active maintenance and wouldn't have an opt-out applied. So it does make sense for FLoC to be opt-in in general.

[lbdvt]

"Ensuring that clustering does not reveal sensitive information" (one way to do it by excluding pages with sensitive content from the Floc building algorithm) and "Ensuring a Publisher can choose whether its sites can be used to build Floc or not" are two valid use cases, but which should be kept separate IMO.

[dmarti]

Added a related issue to capture and expand on the useful opt-in comment from @jkarlin .

[dmarti]

Will a site opt-out that does not require setting an HTTP header be available before the Origin Trial begins?

I can update #47 as needed.

[michaelkleber]

No — adding another opt-out mechanism will require new code, but we're already code complete on the first origin trial.

[TheMaskMaker]

I think, like in the fledge proposal, having publisher controls on interest-group joining could circumvent this problem. This could be a worth-while add to floc.

[ghost]

No — adding another opt-out mechanism will require new code, but we're already code complete on the first origin trial.

oh well. you've created this issue. because of your lack of foresight and planning, you've removed consent from millions of people. if people ask for their consent to be respected, an answer of "we're in a feature freeze" is not only inappropriate, but actually really, really gross

you have an ethical responsibility here and you're refusing to take care of that responsibility because, what, it doesn't fit your release cycle?

editing because i don't want to pong everyone again, but i wanted to add something: i run a few sites on shared hosting (gitlab) that have a target audience of a niche group of marginalized people - most of whom aren't technically inclined enough to understand floc, or how to mitigate it. i'm concerned that my sites might be able to deanonymize people on this marginalized group; the safest course of action is to disable floc on my site, since i frankly don't trust anyone to get floc right enough to avoid things like that. managing my own hosting isn't doable for me currently, so i have no way to protect the people visiting my site

and that's not even getting into how fucked up it is that people are going to be using floc to profit indirectly off of my work, that i create and maintain for free. work that uses shared hosting because i currently can't afford anything else. this isn't just unethical, it's tantamount to theft for someone to profit off of my work in a way that i can't even opt out of because it was too much of an inconvenience for someone to give me a usable off switch

honestly could you imagine going to your boss or your professor and saying "yeah, i'm more than able to do the work you requested, but i don't really feel like it"?
you have become bartleby the scrivener, except instead of inconveniencing one person, you're violating the consent of the entire internet. i'm sure melville would be proud