I realize it's supremely depressing that we have to turn a system long based on trust into Fort Knox. But the alternative will be dystopian. Nothing gold can stay.

And - finally - yes I know this will be expensive to set up and run, but there is no way that the unfettered leakage of rubbish into our scholarly literature is not going to cost us more in the long run. Remember, this is the fountain of knowledge from which the LLMs now drink.

I'm also not suggesting any identity verification system we construct will be fool proof - people will find a way to game/hack anything. And there is no better testimony to our shitty built in incentive system... Goodhart's Law playing out ad absurdum.

I am 100% sure I have suggested nothing revolutionary or novel here - and certainly nothing that @orcid.org hasn't already considered/thought through. (And I hope the experts will chime in!). I'm not suggesting this is simple. I realize it's going to take a huge amount of work to implement.

You could even imagine tiers of verification if people can provide multiple forms. This info could be passed from journal/preprint server submission systems through ORCID's API. That can present a hard barrier to entry or simply carry through to the author list as different colored ORCID badges.

There could be suitable proxy verification through some robust credentialing org: Publons, WoS or - dare I suggest - actual ID documentation through upload of a government issued ID processed through trusted third party, as is surely already done by some funding agencies.

"Ok, but not every scientist is associated with an institution". Right - we will need other modes of verification. We can think about verification that is grants-based, publication-based (robust prior publication record), community-based (some number of other verified authors can vouch).

"But there are probably 10,000 institutions and people move around". Yes - there will need to be some kind of a registry and trusted appointees within the institutions. And perhaps there will need to be periodic re-verification and an audit trail.

but we have gotten to a point where an imperfect solution must be better than the nothing we have now. Requiring all authors on a paper to have an ORCID would be the barest minimum first step. Next, we need actual verification. For the vast majority of authors, that would be through an institution.

The "Paywalls are Evil" slide

In a world where preprints, and not journal articles, are the focus of assessment, there is a significant incentive to ensure one's preprint is complete.