For questions that deal with possible errors in a specific statement of a specified book, article, website, etc. These questions should always include a complete reference.
My problem with this tag: what kind of a classification is it creating? Is it worth having such a classification? I don't see how this classification is useful for any site user in anyway. It does seem to be a list of possible errors in different books. But, this is a very broad list of almost all possible errata from all possible books of all possible fields. Certainly not useful.
Also, I find the usage of this tag can be subjective based on the OP. As a chemistry student, how do I know if what the professors (having more experience than me and having got their proofread by reviewers/publishers) wrote is right or wrong? I can either suspect the book and tag my question erratum, or I trust the book (and describe how I thought the opposite text to be true but it isn't so as said by the book).
Moreover, the usage of this tag has itself been subject to erratum. This question, for example, clearly shows an error in one of the books, there is no subjectivity in it at all. But it did not get tagged erratum. I suspect there might be more such questions missing erratum.
Based on all these evidences, my suggestion would be to delete this tag (go through all questions, remove one by one).