Its probably not that hard to detect. Suddenly, a new domain appears out of nowhere full of articles using specific language (absolutes, 'shocking' terms, etc)

Its probably not that hard to detect. Suddenly, a new domain appears out of nowhere full of articles using specific language (absolutes, 'shocking' terms, etc) and linked to by various low reputation IPs, botnets, etc and you just panda its SEO and go on with your life. There are probably other factors here that make this even easier. These sites use fairly shady advertising networks so that's another factor to tie in. They have English language articles but with the grammatical mistakes a 'real' news outlet wouldn't allow due to being written by algorithms or ESL writers. Fake news detection probably isn't as much about the content but of the methods used to spread it. That's my guess. Unless Google has some incredible AI, they'll just weigh it like they do with other sites abusing SEO. Seems like just a refinement of their panda system focused more on 'news' sites than link farms. The 10,000 staff just are there to handle edge cases and help tweak the system. They're not going to read every moldovian fake news outlet scammer. In fact, considering most of this is autogenerated by the tens of thousands, they simply can't keep up. I kinda of see this like spam filtering for the web. Eventually it becomes economically feasible to generate and promote tens of thousands of fake whatevers (it doesnt have to be news) and Google is just trying to keep up. The larger political issues are moot as it doesn't matter what your bias is in politics. If your search engine keeps feeding you low-information or outright false crap most of the time, you'll think about moving to a different search engine.