How will a system like this handle developing news stories after the fact? E.g. in the Québec mosque shooting there were initially two suspects, and many news articles that had accurate information at the time look incorrect in retrospect. Would those all be marked as BS retroactively by such a system, or is it enough that the news article used the best available facts at the time?
How will it handle granularity? It's very hard to write any news article that's 100% correct about everything. Will the entire article be marked as non-factually accurate due to the slightest mistake, or will they be highlighting specific factual inaccuracies in the content itself?
Whether Google is good or not is irrelevant to whether your particular accusation is baseless.
The perception of whether Google is good or not is all that matters. This is how media and mass psychology works. This isn't about fact-checking stories. It's about using Google's brand to deem specific stories to be real or fake.
But that's not what they're doing! Did you read my whole post? They are not marking articles as checked, they are just tagging them as being Fact Checking articles.
The difference is that it's not a evaluation about whether the content is correct, just about the type of content. It's akin to marking links as "PDF" or "Video".
How will it handle granularity? It's very hard to write any news article that's 100% correct about everything. Will the entire article be marked as non-factually accurate due to the slightest mistake, or will they be highlighting specific factual inaccuracies in the content itself?