Copyright in the Age of Generative AI: who owns content created by machines?

Jelly – Digital Agency

The rise of generative artificial intelligence is reshaping the creative landscape. Text, images, music and even entire videos are now generated from simple prompts. But one question remains without a definitive answer: who is the author of a work created by a machine?

More than a legal debate, this is an ethical and cultural question that cuts across creative boundaries and challenges the business models of the creative industries. And while it all looked like promising innovation at first, it is now clear that the risks of appropriation and of devaluing human authorship are far too real to ignore.

AI as a tool or as an author?

The distinction between using AI as a tool and treating it as the true creator of a work has profound implications. At present, most legal systems, including those of the European Union and the United States, establish that only humans can hold copyright. AI, however sophisticated, has no legal personality and no consciousness. Credit therefore normally falls to the human user: whoever provides the input, adjusts the parameters or selects the final output. But is that degree of control enough to justify authorship? And what if the human contribution is minimal?

Recent cases that raise doubts

As generative AI tools become more accessible, controversial cases are multiplying. These situations not only expose legal gaps but also reveal the growing imbalance between technological innovation and the protection of creative rights.

1. AI-generated books: authorship or operation?

In 2025, reports multiplied of users publishing entire books on Amazon KDP, created with tools such as Sudowrite or Writesonic. With a simple prompt – "write a historical novel with a female protagonist who travels back in time to the 19th century" – hundreds of pages are generated in minutes. Critics argue that this amounts to a kind of "manufactured intellectual authority": the user presents themselves as the author, but the creation is almost entirely algorithmic, underpinned by models trained on decades of literature written by others. AI learns to write by reading the real authors, and never credits them.

2. The music industry and the explosion of deepfake tracks

On platforms such as TikTok and Spotify, AI-generated tracks using imitations of real artists' voices have begun to appear. In 2024, the song "Heart on My Sleeve" went viral by simulating Drake and The Weeknd, and was swiftly taken down following complaints from Universal Music.

In 2025, tools such as Suno AI and Udio made it possible to create entire tracks with replicated vocal identities. This has created a new legal dilemma: is a voice subject to copyright? How do you protect a sonic identity in a world where anyone can replicate it with software?

3. Fashion, design and algorithmic plagiarism

Visual design is also being deeply affected. In 2025, a Spanish designer saw a textile pattern she had created years earlier reproduced almost in full on a garment created with Midjourney. The brand claimed the design had been "originally generated". AI, as ever, explained nothing.

This case is symptomatic of a wider phenomenon: AI does not create from scratch, it remixes what already exists. And when models are trained on vast, unregulated datasets, we are looking at a system that normalises plagiarism dressed up as efficiency.

These examples show that the problem is not the technology itself, but who controls it and how it is used. Today, the big tech platforms hold the power, exploit creative data without transparency and offer tools that benefit end users while rendering the original authors invisible.

Innovation is being built on layers of other people's unacknowledged creativity. As Carlos A. Scolari put it, "implicit consumers become invisible producers" — and that has never been more literal than it is now.

What is being done?

Several international bodies are trying to keep pace with innovation. The U.S. Copyright Office has already been clear: works created solely by artificial intelligence are not eligible for copyright registration unless there is significant creative input from a human.

In Europe, the European Parliament is debating the AI Act and legislative proposals that include a "mandatory AI training licence". This measure would require platforms to financially compensate authors whose works are used to train generative models – an important step, but still a long way from implementation.

The World Intellectual Property Organization (WIPO) warns of the urgent need to create global mechanisms for recognition and compensation, or risk weakening the entire creative ecosystem in the long run.

Creativity is not at risk, but creative justice is. In an era where AI can generate content indistinguishable from the human, it matters more than ever to define who has the right to authorship, to payment and to recognition.

At Jelly, this is not just a legal issue, it is a strategic one. Because communicating with impact in the digital world of 2025 also means being on the right side of ethical innovation.