Generative AI & Copyright: EU vs US Legal Battle for the Future

The European Union and the United States are taking radically different paths when it comes to copyright and generative AI training.

While the EU’s top court is about to weigh in on the first case of alleged copyright infringement by an AI model, US judges have already started embracing fair use arguments from tech giants. In this episode of Diritto al Digitale, Giulio Coraggio unpacks the landmark Like Company v Google Ireland case and contrasts the EU’s strict framework with the more flexible—yet unpredictable—American approach. What will this mean for platforms like ChatGPT, Gemini, and Claude? Tune in to find out.

You can listen to the episode on Apple Podcasts, Google Podcasts, Spotify, and Audible and below

On a similar topic, you can find the article “AI Act – What Is the Scope of the TDM Copyright Exception?“.

Below is the transcript of the episode:

🎙️ [OPENING]

What happens when an AI model “draws inspiration” from a copyrighted article or photograph to generate something new?
 And what if that creation becomes part of a commercial product?

The line between inspiration, transformation, and copyright infringement is becoming increasingly blurry.
 And now, Europe is being called to draw that line—once and for all.

The Court of Justice of the European Union has received its first-ever case dealing with copyright in the context of generative AI.
It’s a battle that has already begun in the United States, where fair use has triggered a wave of lawsuits, headlines, and controversy.
But will Europe follow America’s path—or take a harder stance?

🎙️ INTRODUCTION

Welcome back to Diritto al Digitale, the podcast where we explore the intersection between law and innovation.
I’m Giulio Coraggio, a technology and data lawyer at the global law firm DLA Piper.
And today, we’re diving into a legal issue that could redefine the development of generative AI across the European Union: the first case before the Court of Justice of the EU (CJEU) concerning copyright protection in the age of AI.

We’ll also look at how the United States is approaching the same issue through the lens of fair use, and what this all means for developers, creators, and AI companies.

🎙️ The CJEU Case: Like Company v Google Ireland

The story begins in Hungary, with an eccentric article published by Like Company on the website balatonkornyeke.hu.
The piece featured unusual claims about Hungarian pop star Kozso, including his alleged plan to release dolphins into Lake Balaton—alongside other details gathered from public sources like newspapers and social media.

Later, a user asked Google’s Gemini chatbot to summarize information about Kozso in Hungarian.
 According to Like Company, Gemini’s response echoed the same phrases found in the original article—without attribution, and without any license to use the copyrighted content.

This led to a lawsuit in Hungary, and now, the Hungarian Supreme Court has referred the case to the CJEU.
 At stake: whether the generation of text by an AI model can amount to copyright infringement—even when it occurs in response to a single user prompt.

The Court has been asked to answer three pivotal questions:

• Does AI-generated content that reproduces protected elements qualify as a communication to the public?
• Can such output be considered unauthorized reproduction, even if not published separately?
• And could this use fall under any exceptions or limitations under EU law?

These questions are not just technical—they’re highly political.
 The CJEU’s ruling will set the rules for large language models (LLMs) like ChatGPT, Gemini, and Claude across the entire EU market.

🎙️ The US Approach: Fair Use Takes the Stage

Meanwhile, across the Atlantic, courts in the United States have already weighed in on similar cases—with very different outcomes.

In two recent rulings, US judges accepted that companies like Meta could invoke the fair use doctrine to justify using copyrighted materials for AI training.

One judge summed it up like this:

“No previous case answers the question whether Meta’s copying was fair use. That question must be answered by flexibly applying the fair use factors.”

That’s the key difference: the US fair use system is flexible, based on four broad factors:

• The transformative purpose of the use.
• The nature of the original work.
• The amount of content copied.
• And the market effect of the use.

American judges have recognized that even large-scale, automated AI training can be considered transformative—and possibly fair use—if it supports technological progress.

However, they’ve also made it clear: if the training harms the original market, copyright holders might still win. So the outcome depends on how the facts play out—case by case.

🎙️ EU vs US: Two Legal Worlds, One AI Race

This legal divide reveals two competing visions for how to regulate AI and copyright.

On one side, the EU enforces strict copyright protection, with a closed list of exceptions. That means generative AI developers might have to license every single piece of content they use to train their models—an enormous challenge.

On the other side, the US provides more flexibility, but also more legal uncertainty. Every case becomes a battle, with costly litigation and no clear precedents.

And now that the AI Act has been passed in the EU, more questions arise:

• How will developers be required to document their training datasets?
• Will strict data governance standards apply?
• And who bears the responsibility for copyright compliance—the AI provider, or the end user?

These issues will soon come to the forefront.

🎙️ [ONCLUSION

We’re at a turning point.
 The CJEU’s upcoming decision could shape the entire future of AI development in Europe.
 And the line between technological innovation and copyright violation has never been thinner.

➡️ How should we regulate the use of copyrighted materials in AI training?
➡️ Should Europe embrace a stricter model—or take a more flexible American approach?

I’d love to hear your views. Write to me at giulio.coraggio@dlapiper.com.
If you found this episode insightful, subscribe to Diritto al Digitale and activate the bell to stay updated on future releases.
And if you liked the episode, please give us a 5-star review on Apple Podcasts or Spotify.

I am Giulio Coraggio. This is the podcast Diritto al Digitale.
Arrivederci!