> they can skirt the rules by providing "researchers" with anonymized data.

Probably, given the many past cases of academics using this kind of data to demonstrate de-anonymization attacks, with some very special vetting of who the researchers are and where their loyalties lie.

Here's a decent writeup of their first big research-related privacy blunder: https://www.chronicle.com/article/Harvards-Privacy-Meltdown/...