New AI Copyright Rules

Craig: so copyright and ai has always been a little loose. and historically, we weren't sure where we stand when you create something with ai, what is copyright, who owns it, what isn't. and we've got these new rules that have come out now and that have, i don't know if they've cleared it up at all. they've tried to clear it up at all. let's go out here, judd.

Jon: yeah, this is cool. well, not cool, it's interesting. so us copyright office, they've released a landmark report which addresses the copyrightability of works created using generative ai. the report, again, it's just a report, right? it's part of a series on copyright and artificial intelligence. but the thing that is interesting is it concludes that existing copyright law is flexible enough to handle ai generated content without requiring new legislation. but the key finding in this is that copyright protection can extend to works created with ai assistance, but only when there's sufficient human contribution to the expressive elements, which means that you cannot copyright something that is strictly ai written, right? so not surprising. so it emphasizes that purely generated ai material or content where human control over expressive elements, they're essentially saying it's insufficient and cannot be copyrighted. but when ai is used as a tool and a human determines the expressive elements, then the resulting work may be eligible for copyright protection, which has significant implications for creators, businesses, and the ai industry, right? so this could make ai skills a crucial advantage in the creative and intellectual property landscape. by ai skills, i mean is being able to balance between that generation and making sure that there is expressive elements in the writing. what do you think, craig?

Craig: i think we needed to get somewhere and i'm glad we're moving in a direction. but let's just take a step back and understand what ai copyright rules are. can ai be considered an author? who owns the output? the developer, a user or ai itself? and what level of human input qualifies for copyright protection? so those are the basic questions. and now it's what was the wording again? significant input?

Jon: what was it? sufficient human contribution to the expressive elements.

Craig: so prompt engineering alone is insufficient. business must show a chain of creative decisions to secure rights, substantive human edits, refining ai outputs, adding original elements or using interactive feedback loops, copyright applies to modified portions.

Jon: so this is cool. i get where they're going. and i kind of agree because what they're doing is they're just showing that companies may need to document the human involvement in the ai assisted project more carefully to ensure copyright protection, which makes sense, right? i tend to agree with this. it could lead to new roles and processes in the creative industries, obviously making it a crucial skill for creators and copyright lawyers, which is where i saw this going anyway. what about you? did you see it going this way?

Craig: well, honestly, you know, if anybody's listening, that's a startup. this seems like another billion dollar industry right now. you should start thinking about this, how the interface that audits, you have your work people and their ai so you can audit the interaction so you can claim copyright. if you don't have it, there's no way you're going to get it. you're not going to get those chats later. you're not going to get the reasoning. you're not going to get a lot of stuff. they're going to be gone. so there's an interface here if companies want to retain ip. so there's an ai interface and an ai automation agent and some software tools here to track that, log it, and then send it off to your loyal legal department and say, hey, is there stuff we can copyright in this? and all of a sudden you can do things that maybe you weren't able to do before. yeah. now again, i'm a fan of people owning their work. yeah. but i'm not a fan of corporations going copyright law, copyright law, copyright law, copyright law. but this seems to open up that second option.

Jon: yeah, i agree. yeah. but again, this is needed and this isn't going to be the end of this discussion. this isn't gonna be the end all be all. like, this is just a first, i would say, a first kind of draft or a first attempt at this, but it i think it's a decent first attempt.

Craig: yeah. we're stepping in the right direction. there's a great article on torontostarts dot com where you can find the show notes for all of these podcasts where i have all this broken down. we talk about the global perspective. law? what's uk law? what's japan law? what's china law? china requires clear human oversight for ai content protection. us says strict human authorship requirement. the twenty twenty five ruling tightens its compliance. uk allows copyright for computer generated work with no human author owned by the person arranging the creation.

Jon: sorry, is that uk or eu?

Craig: uk. yeah, not surprised. japan grants copyright if ai output reflects human intellectual contribution. so to protect the ai generated content legally, you need to document the human input. this is what i'm talking about when i said there's an opportunity here for corporations and software and somebody to jump on this, to track edits, prompts and creative decision. like on this one, i've got to use tools like programmer, which is our automation and ai tool website. there's a call out in this one for jumping over there to automate your audits. use disclaimers to close ai involvement in public facing content, leverage hybrid workflows, combine ai tools with human designers, writers, and consult legal experts to review contracts and ip strategies to align with this new twenty twenty five strategy. it's an ever changing thing, but there's things you can do right now to start getting on the right side of this and making sure that you can copyright your stuff. yeah. and with the ever changing landscape of what's going on with the media and using ai for film and creation of content and that kind of stuff, this is becoming more and more important. what do you think, john? do you think we're just at the verge of where we need to go? are we halfway through this journey of getting a decent copyright law? or are we just, you know,

Jon: i think, i think we're getting there. i think this is a step in the right direction. unlike, the uk's trying to get encrypted data, they're going in the wrong direction. but i think this is a great step in the right direction. i do agree. there needs to be some sort of human input if you're trying to copyright something, because then it's just a free for all.

Craig: agreed. and so we'd love to hear your feedback out there. what does cop these ai copyright rules mean to you? there's a huge anti art, anti, you know, well, ai generated images are theft, stuff going on out there. this kind of changes all of that and says, yeah, no, if you've given enough input, actually, you can, that's actually your copyrighted art or material or whatever. we'd love to hear what your feedback in on that. that's great? we're heading in the right direction or oh my god, that's that poor artists, poor people who put all the work into doing the original training of these models, meaning all the artists and content that filled into that that be able to produce the content they're getting screwed over. and when things need to change, let us know what you think because i'd love to hear your feedback.