As reported in the original ORCID PR <https://github.com/DSpace/DSpace/pull/612#issuecomment-59732750>:
My old comment about ORCID havig a separate column in the CSV was addressed, now the GUID is used as the authority string, thus preserving backward compatibility of the CSV. The documentation still has the old CSV screenshot. However, how is CSV import of ORCIDs going to work? I don't see a way how you could just list a new ORCID in the CSV and have DSpace create an authority index document for it (or look it up in existing authority records, for that matter).