LeafOfFaith
Well-known member
For @CH1 since I promised.
I don't know if I see much of a difference between viewing and scraping. I look at some contract, and I go hey, that was smart, I should adopt some version of that in the contracts that I prepare from now on. Same for any artist looking at another creator's work. We've effectively scraped - or, in the ancient vernacular, learned - sometimes willingly and sometimes without even really realizing, in order to repurpose what we saw into something new and better. The AI is doing the same thing, and it leaves behind digital traces of what it accessed, no different than we do these days via browser history, purchases, or, back in the day, library cards. The whole essence of AI is that it learns just like us, and creates just like we do. There's isn't a material difference in the process really, when you strip it down.
I don't see the AI violating any copyright any more than we do. It views and understands something based on search terms (at its most basic level). It's really just googling stuff. We google stuff to learn, research, experience. We also turn on Netflix and watch something that interests us and inspires us to do something. We don't violate the copyright by doing what we do, but the AI violates the copyright by reading text and viewing images available to any human who wants it?
No one is directing us but us (let's leave religious and fatalistic notions out of it), but we are directing the AI. What does the AI do if we don't come up with the idea to instruct it? Nothing. And when we instruct it, we're coming up with original thought and putting tech into action. Obviously, we can't tell the AI to write an original Batman story and publish it online - that's clear infringement of copyrighted IP. But if we tell it to write a story about an orphan boy who suffers great tragedy as a child and then devotes his life to fighting crime as a superhero, and it spits out a new superhero for us, how is that any different than Stan Lee certainly knowing about Batman or Superman's origin story and then coming up with Spiderman? If we tell the AI to write a sad poem, and it scours the internet for sad poems and writes one, is that so different than someone reading poetry all their lives, studying it in school, being familiar with all the great works, and then creating their own? Of course, the counterargument is that the AI is lifting, essentially plagiarizing, but is it doing it any more than a human does in drawing on everything he ever consumed and experienced when creating something? We have styles in every genre - music, movies, books, houses/architecture, etc., and they're "styles" because people liked what someone else did before them, and replicated it in a different way, en masse. Infringement, or the ultimate form of flattery?
Is the person who directs the AI any different than the "person", say Marvel Comics, when it pays the writer, artist, inker, letterer, editor, etc., in creating the comic book? Marvel owns copyright, even though Perlmutter or Iger or whoever weren't even aware of what the workers were doing. The person directs the AI to create something, and it comes up with something original, even though it's an amalgamation of many other works. But that's what I'm getting at here, is that there's a case to be made that the AI process is just using a tool (ChatGPT) to streamline the creation experience by drawing on everything relevant it can find, an advanced form of using the established tool (the human brain) to do the exact same thing.
The practical inevitability of it also is that people WILL use the AI to create a draft, a prototype, a starting (or middle) point for something they want to do. And it will not be detectable because they'll have altered it to fit their purposes and conform to their vision. Short of confiscating hard drives with search warrants to see what they accessed, how is anyone going to prove that the script someone submitted wasn't initially a creation of the AI process that they then tailored into something else? It's here to stay, and it commits the perfect crime almost every time by not leaving behind any fingerprints in most contexts. So, if that's the case, how do we say that something created by AI is not going to be copyrightable? We'll never know what was or wasn't created by AI unless the person tells us. So, fine, the person who tells us they used AI to create something loses the ability to copyright, but what about everyone else who isn't so open about it?
By the way, this is mostly just off the cuff musings coming from my personal philosophical and practical angle. It's a major problem though. I'm not a fan of tech creating IP with a simple one line command. It will produce better results than probably 99% of people every time. I'm also not a fan of self-checkout stands at the supermarket or self driving delivery trucks or little mobile mini pizza delivery bots. I genuinely don't know what people are going to do for work in the rapidly approaching future, and I'm very concerned about the crime, depression, suicide, malaise, and lifelessness that will probably pervade society. But I don't think legislating against advancement will ever really be an effective approach.
this part gets the like. 
