How is one to know that Sweeney, T isn't named Thomas or Theresa? Todd G isn't Todd Gordon?

For years my team had been working on a system [[edit: for our own outside Zotero-facilitated database] that assigns probabilities to name matches using the presence of similar co-authors, date-range of publication dates, and (although not available in Zotero) author's institutional affiliation and ORCID identifier. (We get the affiliation and ORCID by using the Zotero-captured PMID and doing our own polling of the NLM database.) This, at best, helps with human decisions with manual name consolidation.

You can at least identify instances of the same author formatted differently (and other name formatting problems) from an alphabetical list of all authors in your library (but fixing them is up to you). To get that list, run the following code under Tools\Developer\Run Javascript:

var rows = await Zotero.DB.queryAsync("SELECT TRIM(lastName || ', ' || firstName) AS author, COUNT(*) AS num FROM items JOIN itemCreators USING (itemID) JOIN creators USING (creatorID) WHERE libraryID=1 GROUP BY author ORDER BY lastName ASC");
var lines = [];
for (let row of rows) {
lines.push(row.author + '\t' + row.num);
}
return lines.join('\n');

It's always a question with a reference library as to how much time you spend cleaning it up (which can take a long time), or do you just wait until problems become apparent when citing, and only fix authors' names then.

Also, a lesson for young authors: make sure that every time your name appears in the author list on a paper, it is in *exactly* the same format (eg first name spelt out, one middle initial). Journals don't commonly impose the formatting of names in my experience. So where the same author's name appears in different ways across their published work, it's often because they let that happen.