[emil-lp]

One could require DOIs for each reference. That's both realistic to achieve and easy to verify.

Although then why not just cite existing papers for bogus reasons?