Which is an interesting point. Sometimes, especially in the areas of science, the authoritative texts are not readable by the general public, and the journalistic attempts to render them readable often end up being completely non-authoritative, though they come from "authoritative" publications.
Thus, we get headlines like "HRT causes cancer" (sorry, it's more nuanced than that) and "WHI study shows calcium and Vitamin D don't prevent fractures" (well, no, what this deeply flawed part of the study showed is that there was a calcium effect even though they completely failed to control who got how much calcium, and it supported what nutrition researchers have known for 10 years about Vitamin D: 400 IU is too small a dose to bother with for most of our D-deficient population).
The day that Google can actually determine that a newspaper's garbled regurgitation of a scientific study is not authoritative, that'll be some darn slick artificial intelligence they got there :-).
OTOH, it can be entertaining reading Google news when a new scientific study comes out -- sometimes G will put exactly contradictory headlines about the study right next to each other.