How were the polls fake when they were basically right? The state-polls for Wisconsin / Michigan in the end were wrong, but those states in the end were probably not polled enough to capture the volatility of unprecedented events like the Comey letter.
And FYI national polls predicted a 1-4% popular vote win, those where pretty much on the money.
What polls were "basically right" in the month, weeks, and days leading up to the election? I recall most of them being very off the mark because of their methodology. There were a ton of polls that consistently oversampled Democratic voters, women, and other groups to show results that favored Hillary. This is what made them fake polls. They were being conducted to produce a result that would reinforce a narrative.
Most Fake News (NBC, CBS, NYT, etc.) headlined these fake polls while not reporting sampling to their viewers. There's a reason almost all Fake News casters like Maddow were shocked on election night. They bought into their own lies and selective reporting. I learned of the sampling methodology from some of the very sites that Google now reports as "fake" and from Youtubers they have demonetized.
"Polls are used to influence public opinion, not reflect it."
What we're seeing is Google trying to control the narrative through search and monetization now. This is a more-extreme extension of the crap they pulled during the election (where Google search-complete consistently suggested positive things for Hillary and negative things for Trump).
Reminder of fake polls pushed by fake news. Almost all of the establishment media reported polls like this leading up to the election. They traded all credibility to shill for their candidate. But Google did, too, so such sites are still considered "news" by Google's standards.
And FYI national polls predicted a 1-4% popular vote win, those where pretty much on the money.