Is it legal for AI companies to train their models on books, songs, or articles without asking artists’ permission? This is a question I’ve tried to answer previously in this podcast. But I’m revisiting it because a new court ruling in the US just gave us a big clue, and it’s one that content creators need to know about, myself included.
Hello and welcome back to the AI SIX podcast. I’m Joanne Sweeney, and I started this podcast to help you accelerate your AI knowledge, skills, and confidence, just six minutes at a time.
So what’s the story?
The lawsuit: Kadrey v. Meta ended last week. This is the lawsuit where 13 authors, including big names like Sarah Silverman and Ta-Nehisi Coates, took Meta to court, claiming the company used their books without permission to train AI models.
Here’s an article from Wired.
The authors said Meta scraped their copyrighted books from online sites and fed them into its AI systems, all without paying or asking. They claimed this was copyright infringement and that it hurt their ability to earn from their work.
But here’s where it gets interesting. The judge, Vince Chhabria, wasn’t convinced. He said the authors couldn’t show real harm. Could they prove that Meta’s AI was spitting out content that actually competed with their books? Not really. And while the authors focused on the “piracy” side, how the books were obtained, the judge said the real issue was whether Meta’s use counted as copyright infringement under U.S. law.
Meta argued that using the books to train AI was fair use, a legal concept that allows copyrighted content to be used without permission in some cases, like research or creating something new. The judge seemed open to that, especially since there was no clear evidence that it hurt book sales.
Outcome
In the end, the judge dismissed the case. He said the authors made the wrong arguments and didn’t provide enough proof of harm. But he also made it clear this ruling only applies to this group of authors. Another case, with stronger evidence, could go differently.
Future Considerations
So what does this mean for creators and AI? In the U.S., courts are leaning toward allowing AI training under fair use, as long as there’s no obvious damage to the market for the original work. But in Europe, the rules are stricter. Creators can opt out of having their work used for AI training, and developers have to be transparent about what data they use.
The result? A divided legal landscape. AI developers might find the U.S. more flexible for now, but they’ll have to follow much tighter rules in the EU.
Have there been any such cases in the EU?
As an EU-based author, I’m curious as to how this stacks up under our copyright laws. But part of me believes it’s a horse-bolted moment. The models are already trained, AI is bigger than us (sad I know). Here’s the situation in the EU.
| Case/Country | Issue | Outcome/Status |
| Like Company v. Google (EU-wide) | AI training/outputs and copyright | Referred to CJEU; pending decision. Read more. |
| Germany (Hamburg) | TDM exception for AI training | TDM exception applies to research use. Read more. |
| Germany (LAION case) | Photographer vs. AI dataset for training | TDM exception applies; commercial use can opt out. Read more. |
| France (Meta case) | AI training using copyrighted works | Legal action filed; case ongoing. Read more. |
| Czech Republic | Copyright in AI-generated works | No copyright protection unless human originality. Read more. |
The potentially most impactful case for the EU’s AI and copyright landscape is Like Company v. Google Ireland (C-250/25), currently before the Court of Justice of the European Union (CJEU). In this case, Hungarian news publisher Like Company alleges that Google’s AI chatbot Gemini reproduced and made available substantial parts of its protected news articles without consent, arguing both the training and outputs of large language models (LLMs) constitute copyright infringement.
The CJEU is being asked to clarify whether training LLMs on copyrighted content and their resulting outputs amount to acts of reproduction or communication to the public under EU law, and whether the text and data mining (TDM) exception applies to such practices. The court’s ruling will set a crucial precedent for how AI developers and publishers must navigate copyright compliance and the scope of rights in the digital age
Back in the US, Kadrey v. Meta is a win for AI developers in the U.S., for now. But if you’re a creator or working in marketing, publishing, or media, don’t switch off. The legal debate is far from over, and in Europe, the tide may be turning the other way.
About the Podcast Host
Joanne Sweeney is a specialist in public-sector marketing and digital communications. She’s the author of Public Sector Marketing Pro and the creator of the Professional Diploma in AI for Public Sector Marketing. With years of experience in training, consulting, and advocating for digital transformation, Joanne provides insight and guidance for government agencies navigating the digital landscape. She is also the host of the award-winning Public Sector Marketing Show podcast.
