THE INSPIRATION PLACE PODCAST
Miriam Schulman: AI continues to make headlines. Reuters reported a proposed class action against Adobe, an author alleging Adobe misused copyrighted books to train its models without permission. So we need to address how you can protect yourself from being ripped off by the robots.
Speaker 2: It’s The Inspiration Place podcast with artist Miriam Schulman. Welcome to the Inspiration Place podcast, an art World insider podcast for artists by an artist, where each week we go behind the scenes to uncover the perspiration and inspiration behind the art. And now your host, Miriam Schulman.
Miriam Schulman: Hello there! It’s Miriam Schulman, your creator of inspiration. And welcome to the Inspiration Place. Here’s what we’re going to cover. What’s changing in the courts and policy in plain English. What it means for working artists. In other words, your portfolio, licensing commissions, reproduction rights, how to AI proof your value, provenance process, relationships and credibility markers, and a practical action list for artists contracts, website language image handling, watermarking. How to respond when someone clones your style. And finally, we’re going to cover what it means for writers and content creators. So books, blogs, newsletters, courses, scripts, who owns what, what can be licensed, and what to do when your words or your voice gets remixed. But before we get into the how tos, I want to ground this in reality because this isn’t a Twitter argument anymore. This isn’t artists versus tech bros. As a vibes based debate, this has now moved into the grown up room. Lawsuits, settlements, judges actually looking at what these models were trained on and what they spit back and deciding who might owe who money. So this is the headline that made me say, oh, we need to have an episode on this on December 17th, 2025. Reuters reported a proposed class action against Adobe, an author alleging Adobe misused copyrighted books to train its slim LLM language models without permission. So let’s let that sink in for a moment. This is not some sketchy startup in a hoodie.
Miriam Schulman: This is a mainstream creative company, one that you and I have probably both paid. The kind of company that creatives assumed was on our side, on our team. But Adobe is not alone. We just saw a massive $1.5 billion settlement. In a similar fight, authors alleging anthropic used pirated books for AI training. And yes, that was a Bubba with a B. So if you’re listening and thinking, Miriam, I’m a working artist, I’m a mid-career painter, I’m a photographer, I’m a children’s book writer. I’m not Stephen King. Listen to me. This is not about fame. This is about rights and leverage. And when big companies get sued, what happens next? Policies change, contracts change, platforms change. And suddenly you’re agreeing to something you didn’t even realize you were agreeing to. Also, this matters for both writers and visual artists. Courts are paying attention not only to training but to outputs. What the AI produces. Reuters reported that in late October 2025, a judge let author’s claims proceed based on allegedly infringing outputs, meaning the argument isn’t only you trained on my work, but also you generated something very similar to my work. And then there’s the third lane. We have to talk about ownership, because even if you use AI in your process, the US Copyright Office has been clear. Copyright protection hinges on human authorship. Ai only output is not protected, and prompting alone isn’t enough.
Miriam Schulman: Meaning you prompted by yourself. That’s not enough. And that has major implications for writers and visual artists who think, oh, cool, I’ll just generate something and sell it. So that applies to people like me. If I’m using AI to generate maybe parts of this podcast, or it might affect you if maybe you’re using AI to generate an image. So that’s why we’re doing this now, because the question isn’t is AI good or bad? The question is who gets paid, who gets protected, and who gets steamrolled while everyone else is arguing online? Now that we’ve established why now, I want to give you a simple framework so you can follow every headline without your eyes glazing over. Because once you see this, you’re going to start noticing it everywhere. There are three battlefields in the AI copyright wars, and every lawsuit, every policy memo every. We’ve updated our terms of service announcement. You get those emails that you ignore. It’s going to basically be one of these three. First battlefield training data. So translation, did they have the right to use your work to train the model in the first place. And this is the where did the data come from question. Was it licensed. Was it scraped. Was it bought. Was it pirated. And this is where I can give you some personal context. So Harpercollins asked me to sign something. I think this was about a year ago that would allow my book, Artpreneur, to be used to train AI data.
Miriam Schulman: And I’m going to be honest, I didn’t feel like I had enough clarity on what it would actually be used for. So the honorarium was $2,000, which, by the way, I would have to give some of that to my agent. And for the scope of what they were asking, I didn’t feel comfortable. Now, sometimes I second guess that decision because there was a possible or there is a possible upside. If AI is trained on your work, it could earmark you as maybe an authority, or it could make your name show up more in AI answers. Maybe. Maybe. But the problem is, the maybe is not a business term. So if I don’t know the use case, I can’t price the risk. I can’t price the upside. And I’m not in the habit of handing over my intellectual property just because somebody waved some vague promise at me and called it innovation. So when you hear training data, think permission licensing, sourcing, that’s the first battlefield. Okay. Now we’re going to talk about the second battlefield. This is regurgitation. In other words to similar. This is where a lot of creatives get tripped up because they think the whole debate is only about training. But even if some training ends up being treated as fair use in some contexts, the outputs can still create liability, because at the end of the day, copyright infringement is about copying protected expression, and this is especially for writers.
Miriam Schulman: This is the scary part. Text that is outputted from AI can be verbatim or close enough that it’s no longer inspired by. It’s basically you stole my paragraph. Reuters reported that on October 28th, 2025, a New York federal judge refused to dismiss authors claims that ChatGPT outputs were substantially similar to their copyrighted books, meaning that output theory is not getting laughed out of court. Judges are taking this very seriously. So when you see an AI company say, oh, we didn’t store your book or we didn’t scrape your site, remember, that might not even be the only question. So with this battlefield is all about what is the machine producing and does it cross the line into infringement? You know what? Posting every day on social media is not going to magically sell your art. Collectors don’t come from chasing likes. They come from real connections, smart strategies and showing up with intention. Now if you want to learn exactly how to do that, that’s what I teach you inside my free masterclass to watch it on demand. Go to SchulmanArt.com/sell-more-art That’s SchulmanArt.com/sell-more-art. The third battlefield is ownership, and this one is a little sneaky because it impacts not only how you define your work, but how you sell it, license it, and build assets for your business. So the question here is if you used AI, can you copyright the result? The US Copyright Office this is a report from January 17th, 2025 is very clear on the core principle.
Miriam Schulman: Copyright requires human authorship. Materials generated wholly by AI is not copyrightable. In other words, you cannot produce a book by AI. Well, you can, but you can’t get a copyright if you do it that way. Material generated wholly by AI is not copyrightable, and prompting by yourself is not enough. So here’s an important nuance. Ai can still be used like a tool if there’s meaningful human creative contribution. Selection. Arrangement. Editing. Modification. Human authored material inside the final work. These human contributions can be protected. So if you’re a visual artist using AI as a starting point and then you paint over it, collage it, radically transform it, that’s a different conversation than I typed a prompt and printed whatever came out of it. And I know that there are digital AI artists who are doing just that, because they come to me wondering if they can make a business out of it. And then for writers, if you had AI generated draft and you rewrite it and you restructure it and you inject your voice and your stories and your examples and your narrative architecture, again, different conversation than publish whatever the bot spits out. So this battlefield is what rights do you actually own in the work you’re producing and selling? If you used AI.
Miriam Schulman: So those are the three battlefields. Let’s recap them again so they stick. Inputs. Outputs. Ownership. So inputs training data. Did they have the right to use your work to train Outputs. Is the model producing something that’s too close to something that’s already protected? And the last thing is ownership. Can you protect what you make with AI? And what part is actually yours? Okay. Now that you’ve got the map, we can talk about what all this means for working artists and writers, not the tech pundits, especially when it comes to licensing commissions, reproduction rights, and protecting your voice and style. So let’s get into what this actually means. If you’re, first of all, a visual artist like a painter, a photographer or illustrator, mixed media, whatever. If you make images, this section is for you. First, we need to separate two things that everyone blends together because it feels like the same violation. There’s style theft versus copyright theft because in the style of is an emotional gut punch. That’s the part that makes you want to throw your phone across the room. Unfortunately, in court, feelings don’t get you paid, claims do. So the lawsuits have to translate style theft rage into recognizable legal buckets like copying derivative works or even stuff like removed copyright management information, which is basically, did your name and attribution and or metadata get stripped out along the way? That’s why the visual artist lawsuits are so important as bellwethers, because they show what arguments are going to survive and what gets tossed.
Miriam Schulman: Now here’s the second big shift, and this is where I really want you to lean in, because this is actually good news for working artists. Value is moving toward proof, proof and provenance. Meaning in a world where anybody can generate a pretty picture. The people who are serious buyers, the collectors, the design clients, the licensing clients, they’re going to start caring more about receipts. In other words, process receipts. They want to see behind the scenes. They want to see your sketches. They want to see your studio photos. They want to see time lapse. They want to see certificates of authenticity. So all those assets that you thought you were just creating for social media, no, they are now very valuable to you. This is something you can use in court. These are edition logs. This is the story of the piece where it’s been shown why you made it. And I know some of you just rolled your eyes like, oh my God, Miriam, do not make me do more content. I’m not talking about dancing on reels. I’m not talking about even speaking to a camera. I’m talking about Documentation that creates trust, because trust is going to be the premium currency in a noisy market. Okay, here’s the third thing a reality check.
Miriam Schulman: A lot of people are walking around saying, oh, just opt out. And I’m like, excuse me? Opt out. Where exactly? Because opt out is not universal. And here’s a very concrete example. Adobe’s own help documentation for Adobe Stock contributors says they currently do not allow an opt out preference for Adobe Stock content being used for AI or machine learning training. That’s right. If you’ve uploaded images to Adobe Stock, ding ding ding. This applies to you. So if your strategy is I’ll just click a magic button. I need you to understand on some platforms, that button does not exist. Now, what you can do without spiraling and without turning your life into a bunker. You can take some practical steps that are sane and not paranoid. For example, on some platforms there are no AI tags and meta tags. Artstation is an example. If you tag your projects as no AI, Artstation says it will automatically apply a no ai HTML meta tag, and they said their terms prohibit using no AI tagged content to train AI generators. But is that a magic shield? No, but it is a signal. It is documentation. You can create a paper trail that says I didn’t consent to this. And finally, optional because I’m not here to sell you tech tools, but you should know this exists. There are research based tools like glaze that’s G, A, Z, E, and nightshade, and they’re both designed to help protect against style mimicry or to poison training when people scrape your images.
Miriam Schulman: And again, it’s not a perfect force field, but it’s part of the current landscape and you should at least know it exists. So if I boil it down for visual artists, it’s not just AI is stealing, it’s how the stealing gets defined. And the winners in this next season are going to be the artists who can do two things at once. One, protect their work where it’s reasonable, and two build value around what a I can’t fake trust providence and human storytelling. How many times have I been saying this in a row? Human storytelling behind the piece. Okay, let’s talk about writers. And I’m including all of you in this category. Because if you write books, of course, if you write essays, of course, if you write Substack. Yes. Blog posts, some artists write blog posts. What about captions? What about podcast scripts? What about your course content? What about your workshop materials? What? Oh, what about your emails? Because for writers, the AI copyright fight has a different texture than it does for visual artists. And here’s why. Text is easier to prove on outputs. With writing, it’s a lot easier to litigate it, because substantial similarity in text can be very straightforward. You can put two passages side by side and say, this is basically verbatim, or this is close enough that it’s clearly derived.
Miriam Schulman: It’s not just the vibe is similar. It’s this sentence matches that sentence, and that’s why you’re seeing author and publisher cases move very quickly compared to some of the style theft arguments in visual art. And just to put a pin in it, Reuters reported that in October 2025, a judge let author’s claims proceed on the theory that ChatGPT outputs were substantially similar to copyrighted books. In other words, the output based argument is alive and well. So writers, you need to understand the fight isn’t only about did they train on my work? It’s also what does the model produce and does it cross a line? And the money signal is screaming. Anthropic case is like a flashing neon sign that says pirated training sets can become massively expensive. So Reuters and AP both described a $1.5 billion settlement involving authors and anthropic over alleged use of pirated books for training, plus a structure where payments are tied to work. So that’s not oops, our bad. No, you can’t do that. This is a new era of money. So if you were thinking, oh, nobody’s going to get compensated or this isn’t just noise, I want you to clock that number because it tells you exactly where this is going. Licensing payouts and contracts. The contracts are about to get renegotiated, whether you like it or not. So this is the third thing. And this is the part where I want you to be alert and not get anxious.
Miriam Schulman: Contracts are about to get renegotiated. Publishers, agents, platforms. Everyone is going to start slipping in AI language. And if you’re a writer and you don’t want to be deciding your boundaries in the moment, under pressure. When someone says, oh, it’s standard, everybody signs, you want to decide now what you’ll allow. Can your work be used for training? Can it be used for summarization? Can it be used to generate a Miriam style version of your writing? And yes, this is coming fast. What about a synthetic audiobook voice? I don’t know. They can clone my voice and Miriam Schulman read this book. I don’t know, maybe. So even if you’re not a quote unquote big author, don’t kid yourself. The clause logic still trickles down. Because once the big contracts set a precedent, smaller contracts are going to copy and paste the same language. So writers, the headline isn’t AI is writing books. That’s not the headline. The headline is your words are assets, and that people who treat them like assets, meaning they know what they’re licensing and what they’re not are going to have leverage in the next chapter of this. There’s one thing I’ve said in the past, and I’ll say it again. The artists who succeed, they aren’t always the most talented. They’re the ones with the chutzpah to pitch their work, show up in person and ask for the sale.
Miriam Schulman: Playing small is not going to fill your studio or your bank account, but courage will learn the exact steps in my free masterclass, which you can watch on demand. Go to SchulmanArt.com/sell-more-art. Now that we covered what it means for visual artists and what it means for writers, we’re going to move into the part you probably showed up for today, and that is how to protect yourself without living in fear. Contracts, website language, image handling and what to do if someone clones your work and your voice. Now we’re at the part where I’m going to hopefully calm you down, because if you’ve been listening to this thinking, oh my God, Mary, I’m so doomed. No, that’s not my point. You’re not doomed. But you do need to get a little more strategic about treating your creative work like it’s the asset that it is. This is not legal advice. I’m not a lawyer, but this is smart business sense. And I’m going to give you what to do now. Checklist things that are practical and not paranoid. So the universal moves for visual artists and writers first register what matters I know, I know, paperwork is boring. I hate paperwork too, but timing and paperwork become leverage when disputes arise. Registration is one of those annoying adult things that can make the difference between I’m mad that they did this and I have options.
Miriam Schulman: Second, keep process evidence for visual artists. That’s layers. Rafael’s dated exports. Studio shots, sketches work in progress. Photos for writers. That’s drafts version history. Outlines. Dated documents. Tracked changes. This is not because you need to build a courtroom case tomorrow. It’s because in the world of synthetic content, proof of authorship becomes much more valuable. And third, update your licensing language for commissions and commercial clients. You need to spell out what they can do with the work and what they cannot do. And by the way, I’ve been doing this for years. There was actually something in my portrait contract where I say that they’re not allowed to reproduce anything without my permission. Part of that was just to protect me in the sense that my reputation, I didn’t want them taking a janky a picture of my art and making prints of it. But yeah, you need to have this language. You’ve always needed to have this language, by the way. So here is exactly what you need to include for 2026 and beyond. Write this down. Got a pen? You’re taking notes. You need to pause this so you can take notes. You might want to do that. You may want to replay this section more than once. Number one no you may not use this work for model training. No you may not put this into a data set. No, you may not use it to generate derivative works through AI.
Miriam Schulman: This is the kind of language that used to sound dramatic and now it’s going to be just normal. Fourth, stop giving away full resolution training ready files by default. This is a big one. A lot of creatives hand over the entire vault because they’re trying to be easy to work with. You can still be easy to work with without being reckless. Deliver what’s needed for the use case. If they need a file size for print, give them the print file. If they need a web file, give them the web file. But you don’t need to automatically hand over the highest resolution. Do whatever you want with it forever version of your art. Okay, okay, that’s the universalist. I’m going to split it once again. So for visual artists specifically, two extra moves for you use platform level no AI labeling wherever it’s available. Some platforms allow you to label work as no AI, so it’s clearly signaled that you do not consent to it being used to train AI generators. Deviant Art, for example, has an official support page showing you how to toggle a no AI label for each artwork. Is that a guarantee? No it isn’t, but it’s a documented statement of your intent and you want documentation. Second, build Providence into your marketing. I know some of you just want to post the final piece and move on. But Providence is where you win.
Miriam Schulman: So yes, now you do want certificates, addition notes, a documented story process, proof where it was shown, why you made it, what materials you used. And this is no longer fluff, it’s no longer optional. This is how you make your work more collectible in a market where cheap imitations are everywhere. And now for writers, specifically writers, I have two moves for you. First, review your current contracts with publishers, agents, and platforms for any language about AI or machine learning, and also also look for silence because silence can be interpreted very broadly. You want to know what you already agreed to and what you have the power to negotiate next time to decide your bright line. Not what you’ll decide under pressure. What you decide now, is it no training period? Is it training? Only with a license and clear compensation? Or is it case by case? There’s no morally perfect answer here. I already shared I had some ambiguity about what I chose to do, but there’s only. Does it match your values? Does it match your risk tolerance and your business goals? And the reason I’m pushing this is because if you don’t decide your line, someone else is going to decide it for you inside a contract that maybe you didn’t read carefully enough. Okay, okay, that’s your checklist. Not fear, not gloom and doom. Just treat your work like an asset. Tighten your boundaries.
Miriam Schulman: Make it easier for the right clients and collectors to say yes because you’re the real thing and you can prove it. All right, before you go, Here’s your next step. Your final step. If this episode made you realize you’ve maybe been a little too loose with your boundaries, your contracts, your licensing language, what you’re handing over to clients, what you’re posting, where good that awareness is, power, and if you want help putting the right words in the right places so you can protect your work and sell it with confidence, I’ve got you. This is part of the one on one work I do with clients inside the Artpreneur messaging intensive involved. We excavate your stories. Those stories that are truly only yours. We talk about your work in a way that builds trust, communicates value, and makes it crystal clear what people are actually buying when they collect from you. Because in an AI saturated world, your superpower is not more content. It’s clear messaging And credibility and consent. So you can find that over at SchulmanArt.com/VAULT. Now, if you love this episode, or you have a friend who you think will benefit from it, share it with them. Share it with an artist or a writer friend who’s been quietly panicking about AI. And don’t let them spiral alone. All right, now go make something only a human can make. And I will see you same time, same place next week. Until then, stay inspired.
Speaker 2: Thank you for listening to the Inspiration Place podcast. Connect with us on Facebook at facebook.com/SchulmanArt, on Instagram at @SchulmanArt, and of course, on SchulmanArt.com.
Subscribe & Review in iTunes
Are you subscribed to my podcast? If you’re not, I want to encourage you to do that today. I don’t want you to miss an episode. I’m adding a bunch of bonus episodes to the mix and if you’re not subscribed there’s a good chance you’ll miss out on those. Click here to subscribe in iTunes!
Now if you’re feeling extra loving, I would be really grateful if you left me a review over on iTunes, too. Those reviews help other people find my podcast and they’re also fun for me to go in and read. Just click here to review, select “Ratings and Reviews” and “Write a Review” and let me know what your favorite part of the podcast is. Thank you!