Listen to Doctor Ben now, children.
One of the jobs of a reviewer is to ascertain that, where claims are made in a manuscript, they are backed by reference to original research that contains evidence to support the claim. Where such evidence is not present- where there may be a hypothesis proposing what has been ascertained or where there is circumstantial evidence to support it- the claim must be challenged, the manuscript corrected prior to publication and the author rebuked. It happens in every manuscript. It has happened to me. If reviewers aren't scrupulous in researching the evidence for such claims then an unfounded assertion slips through into the empirical literature and becomes a foundation from which further assertions can be propagated. In my (limited) experience this is now far more common than it was twenty years ago. It is certainly the case that I see manuscripts published today that contain appalling errors of experimental design and incredibly poorly evidenced assertions which ignore the depth of empirical data and cherry pick sources to fit a preferred narrative. I can do nothing but promise to challenge such unscientific practices wherever I find them. Believe me when I say that when I rule the world it will all be much better.