The problem with that Facebook video is that his experiment is uniquely designed to show the worst scenario of click farms hits.
At higher numbers and with pages that people legitimately like, the network effects of Facebook do win out. Sure, if you pay for promotion the response will include a lot of click farmers - there's really no way for Facebook to stop that. If you post a worthless page, almost nobody will legitimately like it, nobody will post messages, nobody will view the page, etc. - so the real goal of having your page show up in news feeds due to cross popularity won't happen. Sure, it's surprisingly fast to rack up fake likes - but that has more to do with people not really grasping the scope and size of Facebook's worldwide audience.
Look at the likes of some local ads - Acorns investment is most popular in LA. TouchofModern is most popular in New York. It's not like every single advertising campaign leads to Bangkok being your #1 fan. Unless your targeted demographic has nearly zero interest in your stuff.
I think it's less "fraud" than it is a question of value. Garbage clicks should devalue the service, not render it useless or fraudulent. If you spend $1,000 on advertising, generate a million bogus likes, but your return is $10,000 in sales for legitimate interest - who cares about the click farms? It should be priced into the service.
All the companies paying to advertise through Facebook are not dummies fooled by high like counts and no real-world results.