My only issue here is the 100%. Who is going be spending the time making sure within reason the ads being delivered will both be safe and legit. If they are filled with a bunch of scammers trying get peoples private keys, other information, or infected with malware does not sound like a cheap thing to deal with.
The short answer is the protocol itself would address it. There are a few ideas in the post (blacklists, and thresholds) that could be used. The time spent would primarily be everyone in the community agreeing on the protocol, and then the individual websites integrating it to show the ads if they decide to use it.
It would be nice to one day have googles style metrics on our blogs. I think that is the big thing missing here in bringing in advertisement dollars.
The real challenge is tracking the interaction across all the different Steem interfaces.