…
There are errors anywhere that has a significant level of content, Britannica & Wiki included. Just like there are always bugs in any significant code. Nature and similar top-tier outlets are not immune from the vagaries of human nature [pun intended, as always], but there are no better filters currently in existence until AI advances some more.
As they say, 'data' is not the plural of 'anecdote'
What then can you trust?
Well of course. Quoting from your future biography 'The Man Who Fell To Earth 2', Wiki will one day say "Colif said 'I wouldn't trust a study by Nature'"—are we to discount that just because the reference is to a forum? 🙃 The medium does not define the content—eg Zed has established that half the products on AliExpress are not scams [ref: PCG forums].
So if you won't even use a library as reference, what are you left with?
But seriously—did your course go thru the probably millions of errors present in the content of any large library?