Do we all have to register with all these schemes?
> Your correspondent is a member of an Australian scheme that charges royalties for reprints of news and other content and disburses them to creators who register their works.
The LLM scrapers are indiscriminate and, as they don't track how much each of their billions of nadans was affected by each bit of input, can not possibly provide fair attribution for their final output*. So are all the humans going to be compensated for the scraping done *before* this scheme comes into place or is the compensation done by, ooh, percentage of the total material registered by this scheme that has your name on it?
NOTE: the Australian scheme mentioned is clearly aimed at explicit REPRINTS - that is, the item was found by some means (e.g. a search) and then explicitly taken as something to reprint, with the attribution clearly attached. Trying to apply that style of scheme to chat bot results has NOTHING whatsoever to do with LLM scraping of copyright works: it could be applied to the proactive bots that go out and pull another copy of the work, in which case it not an "AI" issue but instead we are back to the old "Google is showing too much of my text in their search results" problem.
I'm all for compensating the creators, btw; it is just that schemes like this all seem to be working from the outside in and don't seem to match reality. But they'll make *some* payouts, especially to all the big players who are already rich enough to pay the lawyers, and because of that will allow the LLM operators to be able to say "no, no, we are paying, everything is fair now" whilst all the poor young artists are still left starving in their garrets, producing genuinely great stuff that won't even be recognised after their deaths because it has been hoovered up by the spiders before anyone else and the serial numbers filed off before you can blink.
PS
If the answer is "yes, of course you have to register with all these schemes to get your due compensation" how much is THAT going to cost all the aspiring new authors? Even if these databases are free to enter (so how are they funded? By taking too much of your compensation?) there are going to lots of them - and nice people who'll be sure to get your work into all of them for a "reasonable" fee.
* even when paraphrasing, or just plain quoting, a recognisable chunk of text: if it spits out "Nobody expects the Spanish Inquisition" did it get that by reading Monty Python scripts or from ingesting El Reg comments over the years? What if the LLM is found to always mention vultures whenever it produces that quote? BTW, the use here of a trivially small quote is simply to avoid (1) having to make this comment stupidly long and (2) to try to get the reader to ponder how often even quite large chunks of material are found outside of their original piece without attribution "because we all know where it came from" even out of context; imagine how much of a work you can ingest just by reading a fan site** that assumes you know the context... So no responses like "that is too short to be worth paying for".
** now you've got the idea, replace "fan site" with "professional platform for discussing serious subject".