AI and the Creative Economy: The Rights of Artists and Creators

This page is a translation of the original Portuguese text, provided for convenience. It deals with Brazilian law. If there is any discrepancy, the Portuguese version prevails. Read the original in Portuguese

Artificial intelligence is transforming the creative economy and raising urgent legal questions about the rights of artists, musicians, writers, and digital content creators.

The Impact of Artificial Intelligence on the Creative Economy

The creative economy moves significant sums in Brazil and around the world, spanning music, visual arts, literature, design, film and video, games, and advertising. Generative AI tools can now produce text, images, music, and video from simple prompts, and they are advancing quickly. The entire creative ecosystem is facing a transformation without precedent. AI models are trained on enormous volumes of data, including works protected by copyright. That practice raises deep questions about whether using those works is lawful, who owns what machines create, and how to compensate the human creators whose work feeds these systems.

The current landscape pits two fundamental values against each other. On one side are technological innovation and the economic potential of AI. On the other are the protection of intellectual property rights and the dignity of creative professionals. This tension is not merely theoretical. Visual artists have already found their works reproduced, remixed, and redistributed by AI tools without any authorization, credit, or payment. Musicians face situations in which algorithms replicate their vocal styles and compositions. Writers discover that their texts are part of the datasets used to train language models. The debate, then, is not only about technology. It goes to the sustainability of creative professions themselves.

In Brazil, the Copyright Act (Law No. 9.610/1998) provides that protection extends to creations of the mind, expressed by any means or fixed in any medium. The current statute was designed for a world in which only human beings could be authors, which leaves significant interpretive gaps when it is applied to AI output. This regulatory lag is not unique to Brazil. Nearly every legal system in the world is debating it.

Copyright and the Training of AI Models

One of the central questions is whether protected works may be used to train artificial intelligence models. Generative AI systems depend on processing vast datasets that often include books, articles, photographs, illustrations, sheet music, and music recordings. The controversy is whether this use infringes copyright or could fall under a limitation or exception in the law.

In Brazilian law, the limitations on copyright are set out in Articles 46 to 48 of Law No. 9.610/1998. The United States relies on the fair use doctrine, which applies more flexible criteria. Brazil instead works with a list of exceptions. Some scholars consider that list non-exhaustive, but the courts read it narrowly. Current law has no specific provision authorizing text and data mining for AI training purposes, which leaves the practice in a zone of considerable legal uncertainty.

The European Union, through the AI Act (the EU’s Artificial Intelligence Regulation) and the Directive on Copyright in the Digital Single Market, has already set clearer parameters. The European directive provides exceptions for text and data mining in scientific research, but it allows rightsholders to object to commercial use of their works for that purpose (an opt-out mechanism). This approach seeks to balance innovation with the protection of creators, and Brazilian lawmakers have been studying it as a possible model.

Protecting the rights of artists and creators in the age of artificial intelligence is not only a legal question. It is an ethical imperative for preserving cultural diversity and the sustainability of creative professions.

In Brazil, Bill No. 2.338/2023, which establishes a regulatory framework for artificial intelligence, touches on copyright only in passing. It does not yet offer a definitive solution to the problem of training models on protected works. The legislative debate has been intense, with input from organizations representing the creative sectors, technology companies, academia, and civil society. The challenge is to build a legal framework that does not stifle innovation but also does not sacrifice creators’ rights in the name of technological progress.

Ownership of Works Generated by Artificial Intelligence

Another fundamental issue is who owns works produced by artificial intelligence systems. When an AI tool generates an image, a text, or a piece of music, who is the author? The software developer? The user who supplied the prompt? The machine itself? Or, in the absence of a human author, does the work simply receive no copyright protection?

Brazilian law, like most Western legal systems, ties authorship to the natural person who creates the intellectual work. Article 11 of Law No. 9.610/1998 is clear that an author is the natural person who creates a literary, artistic, or scientific work. In principle, this definition excludes both legal entities and machines from authorship. Administrative decisions in other countries have already taken this position. The U.S. Copyright Office, for example, has refused to register works generated entirely by AI, recognizing protection only where there is substantial human creative contribution.

This position, however, creates a paradox. If AI-generated works are not protected by copyright, they would immediately fall into the public domain, and anyone could use them freely. In theory, that could benefit the public by expanding the pool of available works. On the other hand, it could discourage investment in developing creative AI tools, since their output would have no legal protection.

A middle-ground approach is gaining traction in the international debate: recognizing different degrees of protection depending on how much humans are involved in the creative process. When a person uses AI as a mere tool (setting aesthetic parameters, selecting results, editing and refining the final product), there would be enough creative contribution to justify copyright protection. When generation is fully automated, with no curation or meaningful human intervention, the work would have no author and therefore no protection. This distinction is conceptually coherent, but applying it case by case presents considerable practical challenges.

Protecting Creators’ Image, Voice, and Style

Beyond copyright in the strict sense, artificial intelligence raises significant concerns in the field of personality rights, especially the protection of artists’ image and voice. Deepfake and voice-cloning tools can reproduce anyone’s appearance and voice with striking accuracy, creating content that simulates performances the original artist never gave.

In Brazil, the protection of a person’s image is guaranteed by the Federal Constitution (Art. 5, items V, X, and XXVIII) and by the Civil Code (Art. 20). Using someone’s image without authorization for commercial purposes, or in a way that harms their honor, reputation, or standing, gives rise to a right to compensation. Scholars and courts treat the voice as an attribute of personality, and this protection extends to it. Using AI to clone a singer’s voice without permission, for example, violates personality rights, even without any specific regulation on artificial intelligence.

Artistic style, by contrast, is more complex. An artist’s style (technique, color palette, way of composing) is generally not protected by copyright, because protection covers the concrete expression of a work, not ideas, methods, or styles. This means that an AI tool that generates images “in the style of” a particular artist, without reproducing specific works, may not infringe copyright in the traditional sense. Still, the question is far from settled. There are solid arguments for seeking protection through other legal avenues, such as unfair competition or unjust enrichment, especially when the style is used commercially and causes economic harm to the imitated artist.

Protection and Self-Regulation Initiatives

In light of these legislative gaps, a number of self-regulation and technology initiatives have emerged to protect creators. Tools such as Glaze and Nightshade, developed by researchers at the University of Chicago, let artists apply imperceptible alterations to their digital images, making it harder for AI models to learn their styles from those works. Creative content platforms have adopted transparency policies requiring AI-generated content to be labeled as such. Collective rights management organizations in several countries are beginning to discuss licensing models for the use of works in AI training, aiming to ensure payment to rightsholders.

The Content Authenticity Initiative (CAI), led by major technology and media companies, proposes technical standards for tracking the provenance of digital content, making it possible to verify whether a work was created by humans, by AI, or by a combination of both. These technological solutions are promising, but they do not replace the need for an adequate legal framework. They can, however, complement it in a meaningful way.

Regulatory Outlook and Paths Forward

The global regulatory landscape is evolving rapidly. The European Union leads with the AI Act, which imposes transparency obligations on generative AI systems, including disclosure of the data used in training. The United States is moving forward with executive orders and state-level legislative initiatives. China has had specific regulations for generative AI in place since 2023. In Brazil, besides Bill 2.338/2023, several other bills are pending that address different aspects of the relationship between AI and copyright.

Some guidelines appear to be converging internationally: transparency about the data used to train AI models; creators’ right to object to the use of their works for that purpose (opt-out); the obligation to label AI-generated content; and the pursuit of fair compensation mechanisms for rightsholders whose works feed these systems. The challenge lies in implementing these guidelines effectively, given the transnational nature of the technology and the differences among legal systems.

For Brazilian artists and creators, some practical measures can already be taken, even without specific regulation. Formally registering works with the competent bodies strengthens proof of authorship. Monitoring how works are used on digital platforms makes it possible to spot potential violations. Joining collective management organizations makes it easier to defend rights at scale. And seeking specialized legal advice is essential to assess each situation and identify the best protection strategies.

We conclude that a balance between technological innovation and the protection of creative rights is both possible and necessary. Artificial intelligence can be an ally to creators, expanding their expressive possibilities and opening new markets. But that will happen sustainably only if artists’ rights are respected, transparency is guaranteed, and fair compensation mechanisms are established. The law has a central role in building this balance, and creators’ active participation in the regulatory debate is indispensable if their voices are to be heard.

This subject is closely tied to generative AI and copyright, a topic we cover in a dedicated article.

This subject is closely tied to data subject rights under the LGPD (Brazil’s General Data Protection Law), a topic we cover in a dedicated article.

As informações deste artigo são de caráter informativo e não substituem consulta jurídica individualizada. Cada caso possui particularidades que exigem análise específica.