There ads would be filterable, so that someone viewing the database could filter based on the same criteria that Facebook allows advertisers to target. So if I search "non-college educated white males over the age of 40 in Midwestern suburbia", I get a complete overview of what these people have been shown. If I come across a nefarious advertiser, I can also filter to see what other campaigns they've run and who they've targeted.
Transparency, solved. Now the responsibility to police the ads becomes somewhat shared and it's easy to imagine watchdog groups monitoring different types of campaigns looking for bad actors.