Hello. I've had two websites contributing to AI learning models, and I'm not sure whether to feel flattered or peeved (because they neither asked me nor paid me). People can search Google's dataset to find out if their website has been trawled by AI. There's a search box right near the bottom of this article: https://www.washingtonpost.com/technology/interactive/2023/ai-chatbot-learning/ I quite like the approach taken by the UK's Society of Authors, which seems to be pragmatic, though possibly also naively optimistic. Their statement on AI from an author's point of view, with a link to a page on practical advice for dealing with publishers, is here: https://www2.societyofauthors.org/where-we-stand/artificial-intelligence/
An interesting take! I never knew about the beginnings of American publishing. I do find myself on the side of authors in this one, though—despite all AI and copyright things soaring right over my head.
My substack audience is much smaller than yours; mine numbers only in the hundreds. But grasping copyright, the concept, was the bulwark I relied on back in 2018 at a blogspot site that expanded my skills by quarter-inch lengths to get to where I am now. With that legal-conceptual wall in place I was able to muster the will and marshall the energy on dozens & dozens of separate occasions for each original issue and original illustration I made. ...Beyond my comment, I found this to be superb, from a writer more in line with your current position: https://www.wired.com/story/confessions-viral-ai-writer-chatgpt/
It's refreshing to see this side of the discussion. Gonna be honest, I'm not particularly phased either. With all the data and privacy abuses over the internet, AI-generated content is the least of my concerns. When creative work is transformed enough to be distinct from the original work, it's not theft. It's transformative work - period. regardless of how the narrative is spun to make it seem that way. And just like with any claim of plagiarism or copyright abuse, it should be taken on a case-by-case basis to determine if the work is truly transformative or if it's thinly-veiled plagiarism.
On the other hand, I can definitely agree with the regulation of AI from disrupting job employment. I also believe that we, in the writing & art world, as a collective have created (or aided) this problem by allowing digital junk food to take over the industry. It's easy to say that in hindsight of course, and it's almost inevitable given the accessibility, democratization, and lack of gatekeepers on the internet and beyond. But understanding the whole of the problem, not just the small bits that trigger us, is how we move forward.
Wow I so don't get this, and I didn't get Doctorrow's memorable comment either. Maybe I need an illustration to understand what you both are saying. And I'm aware of early American newspapers ripping off each other's stories (copy) in plain sight on the regular. I think of Hunter Thompson copying Faulkner word-for-word on the typewriter. He ran Faulkner's work through a human operating system https://www.youtube.com/watch?v=JVcrimSicek It's the humanities. Also, Thompson didn't turn around and cynically sell that piece of typed-up prose. See also: Basquiat's interpretation of Van Gogh's self-portrait.
I share many of the same conflicting thoughts about the limits and complications of copyright and authorial labor, but I am cheering on these lawsuits, even as I have no idea how they're going to turn out, or even ultimately, if I side with the authors affected, assuming I may be one of them.
The reason I cheer the lawsuit is because for better or worse, these suits are a way to hash out the unknowns around these issues and bring some light to the darkness. The simple act of forcing the companies to reveal their training data via the litigation discovery process would be a net good. (Which is why I hope there won't be a quick settlement.) The opportunity to generate more pieces like this one which force us to grapple with the complexities is good.
Emotionally, I'm in the camp of not feeling all that much about the potential taking of my work without permission for these purposes, but it does give me some me some pause that OpenAI is claiming a 90 billion dollar valuation on a product built on this labor.
I can't muster any high dudgeon about this, but I would like to know a lot more about what's going on.
I think deep down it has implications well beyond authorial copyright. It's really about things you generate that could conceivably be used as "data" for the business purposes of these companies. Our personal data has been hoovered up and given to this enterprise for a long time now and no one even pretends that we should get paid for it anymore.
For me, the difference with LLMs is that they produce something very much like what I do with my labor, so that begins to feel troublesome in ways grocery stores tracking my spending habits to feed me coupons doesn't.
But yes, how do I get paid for my stuff so I can live a life of reasonable dignity? If OpenAI needs to be bodychecked by name-brand authors for me to keep on keepin' on, I've gotta support that, right?
Hello. I've had two websites contributing to AI learning models, and I'm not sure whether to feel flattered or peeved (because they neither asked me nor paid me). People can search Google's dataset to find out if their website has been trawled by AI. There's a search box right near the bottom of this article: https://www.washingtonpost.com/technology/interactive/2023/ai-chatbot-learning/ I quite like the approach taken by the UK's Society of Authors, which seems to be pragmatic, though possibly also naively optimistic. Their statement on AI from an author's point of view, with a link to a page on practical advice for dealing with publishers, is here: https://www2.societyofauthors.org/where-we-stand/artificial-intelligence/
An interesting take! I never knew about the beginnings of American publishing. I do find myself on the side of authors in this one, though—despite all AI and copyright things soaring right over my head.
My substack audience is much smaller than yours; mine numbers only in the hundreds. But grasping copyright, the concept, was the bulwark I relied on back in 2018 at a blogspot site that expanded my skills by quarter-inch lengths to get to where I am now. With that legal-conceptual wall in place I was able to muster the will and marshall the energy on dozens & dozens of separate occasions for each original issue and original illustration I made. ...Beyond my comment, I found this to be superb, from a writer more in line with your current position: https://www.wired.com/story/confessions-viral-ai-writer-chatgpt/
It's refreshing to see this side of the discussion. Gonna be honest, I'm not particularly phased either. With all the data and privacy abuses over the internet, AI-generated content is the least of my concerns. When creative work is transformed enough to be distinct from the original work, it's not theft. It's transformative work - period. regardless of how the narrative is spun to make it seem that way. And just like with any claim of plagiarism or copyright abuse, it should be taken on a case-by-case basis to determine if the work is truly transformative or if it's thinly-veiled plagiarism.
On the other hand, I can definitely agree with the regulation of AI from disrupting job employment. I also believe that we, in the writing & art world, as a collective have created (or aided) this problem by allowing digital junk food to take over the industry. It's easy to say that in hindsight of course, and it's almost inevitable given the accessibility, democratization, and lack of gatekeepers on the internet and beyond. But understanding the whole of the problem, not just the small bits that trigger us, is how we move forward.
Wow I so don't get this, and I didn't get Doctorrow's memorable comment either. Maybe I need an illustration to understand what you both are saying. And I'm aware of early American newspapers ripping off each other's stories (copy) in plain sight on the regular. I think of Hunter Thompson copying Faulkner word-for-word on the typewriter. He ran Faulkner's work through a human operating system https://www.youtube.com/watch?v=JVcrimSicek It's the humanities. Also, Thompson didn't turn around and cynically sell that piece of typed-up prose. See also: Basquiat's interpretation of Van Gogh's self-portrait.
I’m all for whatever helps creatives protect their intellectual property. Guardrails, etc.
Have you read Peter Schoppert's Substack? https://substack.com/@aicopyright
Just subscribed! Thanks for the tip
I share many of the same conflicting thoughts about the limits and complications of copyright and authorial labor, but I am cheering on these lawsuits, even as I have no idea how they're going to turn out, or even ultimately, if I side with the authors affected, assuming I may be one of them.
The reason I cheer the lawsuit is because for better or worse, these suits are a way to hash out the unknowns around these issues and bring some light to the darkness. The simple act of forcing the companies to reveal their training data via the litigation discovery process would be a net good. (Which is why I hope there won't be a quick settlement.) The opportunity to generate more pieces like this one which force us to grapple with the complexities is good.
Emotionally, I'm in the camp of not feeling all that much about the potential taking of my work without permission for these purposes, but it does give me some me some pause that OpenAI is claiming a 90 billion dollar valuation on a product built on this labor.
I can't muster any high dudgeon about this, but I would like to know a lot more about what's going on.
Same! I am waiting for that smart person better versed in law and business to help me think it all through. We have to follow the money, right?
I think deep down it has implications well beyond authorial copyright. It's really about things you generate that could conceivably be used as "data" for the business purposes of these companies. Our personal data has been hoovered up and given to this enterprise for a long time now and no one even pretends that we should get paid for it anymore.
For me, the difference with LLMs is that they produce something very much like what I do with my labor, so that begins to feel troublesome in ways grocery stores tracking my spending habits to feed me coupons doesn't.
But yes, how do I get paid for my stuff so I can live a life of reasonable dignity? If OpenAI needs to be bodychecked by name-brand authors for me to keep on keepin' on, I've gotta support that, right?
At any rate it’s not either/or. I’d be so up for someone to try a labor approach to this in addition to a copyright one