Hello, Genius! Welcome back to the Own Your Genius podcast—the space where we encourage you to use your education and experiences to build dope businesses. I’m your host, Brand Attorney LaConya Murray, owner of Off the Mark IP Solutions and founder of MARKEDlegal.
A week or so ago, I came across a post on Threads where the author said they only share certain content online because not enough IP attorneys are talking about how to protect content from AI. Fair enough—I get it. The challenge is that when it comes to AI and copyright, the law is playing catch-up. Courts are trying to apply laws to situations that weren’t even imagined when those laws were written.
Despite that, I believe we can find guidance by remembering why copyright laws exist in the first place: to foster creativity. If people know their work can be copied freely, they’re less likely to invest time, money, and energy into creating. But if the law gives them control, they’re more likely to keep creating—and society benefits from that creativity. So we want to make sure that protection remains strong.
One of the biggest concerns today is that creative work is being used to train AI. That’s why today’s episode dives into a case at the heart of this issue: Kadrey v. Meta. Let’s get into it.
What Is This Lawsuit About?
The case is Kadrey v. Meta—yes, Meta as in Facebook. Thirteen authors filed a lawsuit claiming Meta downloaded their books from unauthorized online libraries and used them to train its AI model, Llama, without permission or payment.
Llama is Meta’s large language model—a type of AI that learns how to write by analyzing massive amounts of text. The better the training data, the better the model performs. That’s why books, which are well-written and structured, are especially valuable. Meta used these books to help Llama generate human-like responses, summaries, and even creative writing.
What Did the Authors Claim?
The authors filed several claims, including:
- Direct copyright infringement: Meta allegedly reproduced their books without permission.
- Vicarious copyright infringement: Meta allegedly benefited financially from the infringement while having the ability to stop it.
- DMCA violation: Meta allegedly removed or ignored copyright management information protected under the Digital Millennium Copyright Act.
- Unfair competition: Meta’s actions allegedly gave it an unfair advantage over competitors who follow copyright rules.
- Unjust enrichment: Meta allegedly profited from the authors’ work without compensating them.
- Negligence: Meta allegedly failed to take reasonable steps to avoid infringing on the authors’ rights.
- Distribution: Later, the authors added a claim that Meta may have re-uploaded the books during the AI training process, potentially sharing them with others.
Why Does This Matter?
Copyright law gives creators a bundle of exclusive rights:
- The right to reproduce their work (make copies)
- The right to distribute it (share or sell)
- The right to create derivative works (adaptations, translations)
- The right to publicly display or perform the work
These rights exist to protect your creative genius and ensure you can profit from your work.
But here’s the twist: the law also allows for fair use, a legal defense that lets others use copyrighted material without permission in certain cases—like education, commentary, or research.
Fair use is judged using four factors:
- Purpose of the use: Is it commercial, educational, or transformative?
- Nature of the work: Is it factual or creative?
- Amount used: How much of the work was copied?
- Market impact: Does the use hurt the creator’s ability to earn money?
What Did the Court Decide?
Most of the claims were dismissed early on, except for direct copyright infringement and distribution. Both parties filed motions for summary judgment to determine whether Meta’s use of the books to train Llama was fair use.
A motion for summary judgment asks the court to decide a case or issue without a full trial, based on undisputed facts. In this case, both sides agreed that Meta downloaded the books and used them to train its AI. The question was whether that use was legal.
The plaintiffs argued:
- Meta copied their books without permission.
- Llama could reproduce snippets of their work.
- Meta’s actions hurt their ability to license their books for AI training.
Meta responded:
- The use was transformative—training AI is different from reading a book.
- Llama couldn’t reproduce enough of the books to matter (no more than 50 words at a time).
- The authors weren’t entitled to control the market for licensing books to AI companies.
Ultimately, the court ruled in Meta’s favor—not because the use was clearly legal, but because the plaintiffs didn’t present strong enough evidence of market harm. The judge emphasized that this ruling only applies to these thirteen authors and doesn’t mean Meta’s actions are lawful across the board.
Where Does the Case Stand Now?
The case isn’t over. The remaining issue is whether Meta distributed the works after downloading them. The plaintiffs argue that Meta may have re-uploaded the books during the training process, which could constitute unauthorized distribution.
The court has scheduled discovery dates to investigate this further.
The Bigger Picture: AI and Copyright Law
Copyright law wasn’t built with AI in mind. That’s why courts, lawmakers, and creators are scrambling to figure out how to apply existing protections to this new technology.
Right now:
- There’s no specific law that says you can’t use copyrighted content to train AI.
- Fair use is the main legal battleground.
- AI companies argue training is transformative and legal.
- Creators argue it’s copying—and harmful to their market.
What You Can Do to Protect Your Genius
Even without a concrete law, you can take steps to protect your work:
- Register your copyrights: It strengthens your legal position and gives you the right to sue.
- Add AI-specific notices: Like the one at the end of How to Train Your Dragon that says the movie can’t be used to train AI. It’s not enforceable on its own, but it signals non-consent.
- Track how your work is used: Tools like Glaze, Nightshade, and Have I Been Trained can help detect AI training and output patterns.
- Join the movement: Creators, publishers, and advocacy groups are pushing for stronger protections. Your voice matters.
Watch what companies like Pixar and Disney are doing—Disney has historically led the charge in copyright advocacy.
Final Thoughts
The intersection of copyright law and AI is anything but settled. Creators have powerful rights, but they must navigate ongoing debates about fair use and market harm. For now, there’s no definitive law stopping AI companies from training on copyrighted content. That’s why vigilance and proactive measures are more important than ever.
And while the jury is still out on what infringement looks like in the age of AI, we’re hosting a Trademark Infringement Workshop on August 19th at 11 AM CT. You’ll learn how to:
- Spot signs of brand identity theft before it spreads
- Understand what trademarks actually protect—and how to claim yours
- Respond strategically when someone copies your name, logo, or signature offer
- Use cease & desist letters and trademark registration to shut it down—legally and confidently
Thank you so much for joining me.
