That's a good system, but I'm not sure it's actually workable at the scale of Twitter or Facebook's ad platforms. That's a a lot of human review.
Using "scale" as an out doesn't fly. Sometimes you have to cut back on growth and/or scale to do the right thing. No human review clearly hasn't worked, so either automate or get humans in there. More likely, start with the latter and replace with the former over time.
Automating can help with the more obvious stuff, but there are many ways it can fail in either direction. For example, failing to detect hate-speech dog whistles, or inadvertently suppressing ads from minorities where the language patterns fall outside of what it sees as "the norm".
Anyway, like I said, I'm not disagreeing with you. These systems definitely need considerable improvement, but doing that isn't as trivial as it may at first seem.
Even if they lose money by manually reviewing a small ad that was flagged as possibly political, it doesn't mean it costs more to review 1/100 low-volume ads. They're also paying to classify useful data to train their model.
For example, Facebook crawls links sent in messenger. [0]
No source, but I believe the major advertising platforms, e.g. Google Ads are doing this today to check for banned content.
Facebook could build a heuristic of banned pages & not allow ads to be published. Humans could manually verify things at the edges.
There is a whole industry called link cloaking, that shows the crawlers a sanitized page, and the human targets the actual landing page.
It's a cat & mouse game between ad publishers & blackhats, often times foreign nationals outside of US jurisdiction.
I'm trying to think through the scenarios where a political campaign might use blackhat techniques to subvert verification crawls, and it's not clear that it would stop all cases, but I believe it would stop most.
Political campaigns could use the Facebook accounts & payment information of foreign nationals to temporarily push ads until they were reported - would the liability of blowback be enough to prevent it?