Earlier quoted context omitted.
> I don't think those technologies are well suited to this problem at all. How so? Deciding whether news is fake or not seems like a textbook application of natural language classification and search.
The only difference between a fake news story and a real news story that you can rely on is that the events in the real news story actually took place. Obviously there is no way to determine that by looking at the text itself. Yes, at present the fake news may be written more colorfully or poorly. Obviously if you filter on that difference, it will go away.
This is IMO a much easier problem than a lot of the NLP work for sentiment analysis and other marketing stuff that is working or being funded today. The problem is not technical, the problem is that fake news is profitable.