Especially since they suddenly become not so sure when talking about feeding things under IP to “AIs”. It seems that when some process is not too open, like dataset collection, people doing it get used to bending laws they themselves rely on.
Actually this should be leveraged.
One approach - IP is solid, so those big companies championing “AIs” will have to pay royalties for everything produced by an “AI” which had been fed something of that IP. That’s just logically a Gordian knot.
Another approach - IP is an artificial concept which is complete bullshit, then “digital piracy” is not a crime, and neither is commercialization of fan works over some IP without paying royalties.
Anything in between would mean that a company has more rights under the law than an individual. Would be a good analogy to cutting that knot IMHO, but a bad outcome.
Especially since they suddenly become not so sure when talking about feeding things under IP to “AIs”. It seems that when some process is not too open, like dataset collection, people doing it get used to bending laws they themselves rely on.
Actually this should be leveraged.
One approach - IP is solid, so those big companies championing “AIs” will have to pay royalties for everything produced by an “AI” which had been fed something of that IP. That’s just logically a Gordian knot.
Another approach - IP is an artificial concept which is complete bullshit, then “digital piracy” is not a crime, and neither is commercialization of fan works over some IP without paying royalties.
Anything in between would mean that a company has more rights under the law than an individual. Would be a good analogy to cutting that knot IMHO, but a bad outcome.