Yeah, this is "extractive AI": take original articles written by humans and churn them into machine-spam. It's inauthentic, but that doesn't matter, only the ad impressions do.
How does the machine know what it's telling us is correct?
This is an extremely complicated philosophical question, and it's not limited to machine-generated text; there's already plenty of human-written blogspam. In order for text to be valuable it has to be trustworthy. That is, we have to be reasonably convinced that the author has made some effort to be accurate, to correspond with some mental model of the world which itself corresponds reasonably well with objective reality.
Otherwise it's just an elaborate generative fiction tool writing fanfiction about tyres to sell tyres.