If the ads are billed per period of time and not per-click or per-view then "fraud" doesn't matter; all that matters is the overall ROI; if the advertiser gets more business than what it cost them to run the ad they will consider it a worthwhile investment, otherwise they will stop advertising there forcing the publisher to adjust their prices and eventually those prices will reflect the true value (as in how much ROI it brings) of the ad spot regardless of any "fraud".

It's not that fraud doesn't matter, it just moves to the publisher because they're the ones that have to track the eyeballs now in order to advertisers to know how good their ad space is.

if you have 3 campaigns with 3 ad networks or ad platforms, how do you measure the ROI of the campaigns without tracking?

How do you know if Facebook or pinterest is better ROI for your marketing budget?

You can do first-party tracking where every ad would link to your website with a different query parameter (https://example.com?campaign_id=XYZ).

You then just count the number of purchases from each campaign, compare that with the amount you spent on that campaign and there's your ROI.

Sorry, I misunderstood your post then. So you are in favor of tracking...

The OP question was about banning all tracking.

If you're not paying for clicks, there's no such thing as "fraud" traffic -- there's just users who convert and buy your product and those who don't.

It doesn’t matter because your traffic is measured by Comscore or Nielsen or some other third party and can’t be faked. (I don’t consider being in a survey group “tracking” in the relevant sense.)