Re: [admin] Consensus Call: Adopting Privacy-Preserving Attribution into PATCG (#26)

Hello all here, I wanted to discuss a few things: 

1. While we welcome the IAB to present the AdMAP proposal to PAT, it is simply not relevant to this discussion. 

Wearing my hat as the contributor from The Washington Post for a moment, I do not believe it is particularly private by most definitions, it is just secure. No system that does one to one ID matching is really "private". AdMAP is a Clean-Room-Based solution, which means it handles one to one mapping of IDs. It cannot run in either a user Do Not Sell state nor in an Opt In region where the user has opted out in my opinion, since it is explicitly transmitting specific user IDs. 

Wearing my co-chair hat I don't see how it could possibly be considered 'more private' than any of the family of private measurement proposals that we've considered, much less the latest one. Considering our goals, our privacy principles, and the charter of the upcoming WG, I don't see how we could actually even pick up work on AdMAP--if the IAB brought it to us for standardization work in this context, which I can't see why they would--nor how it would be appropriate. AdMAP could be considered as potentially a system that could run along side other work this group outputs, or a consumer of the APIs that proposals might produce (though I can't see how with my current knowledge of it). At this point, I'd prefer this thread of conversation to drop discussion of AdMAP, I don't think it is relevant. 

2. In terms of fraud concerns, I see and hear them raised for this proposal. I think we can take up the proposal and prioritize addressing fraud concerns actively. I believe the specific innovations to the previous proposals around the privacy budget and correct structuring of the `impressionSites` field can give us better antifraud results then we have in the current marketplace to be honest. This can further be enhanced on the antifraud front when user agents have both this and other proposals in the antifraud working group online. The discussion around filtering during TPAC gave me a lot of confidence that there are properties that can be further worked on in this proposal to address your antifraud concerns @dmarti, though of course that development will take time and input like yours to make sure we get right. We can also bring online AdMAP discussions or any other anti-fraud measure discussions that participants might decide to leverage to further bring concerns around antifraud to the forefront. I don't think that PPA is lacking in hooks for, as @wbl notes ways for entities to be "applying their own integrity measurements to impressions or reports of conversions".

3. Wearing my co-chair hat I think I feel comfortable in saying that if we adopt the proposal we can have presentations and potentially, if you are willing to lead it @dmarti, an ad-hoc group specifically to make sure that PPA has antifraud measures that satisfy the wider group. 

I don't think there are any other objections to consensus here, but I'd like to make sure you feel like your concerns are addressed and that they will be incorporated into the documents we produce. @dmarti considering the time spent on this conversation thus far, I think could you respond on if you continue to hold up your objection here within the next three days (Thursday afternoon) and after that @seanturner and I will adjudicate on consensus. 

-- 
GitHub Notification of comment by AramZS
Please view or discuss this issue at https://github.com/patcg/admin/issues/26#issuecomment-2427322389 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Monday, 21 October 2024 17:31:24 UTC