fbpx

AI and Copyright: Can Machines Truly Own Creativity?

The evolution of artificial intelligence is changing the landscape of creativity. From generating artwork and composing music to crafting intricate stories, AI is increasingly venturing into what was once considered an exclusively human domain. But as AI-generated works flood the creative market, a pressing question looms: Can these creations be protected by copyright law, and if so, who owns the rights? Is it the AI, the developer, or the person who prompts the AI? Unfortunately, the answer is far from simple and reveals the intricate interplay between creativity, law, and technology.

The Traditional Framework of Copyright Law

Copyright law has long been the guardian of creative expression, designed to protect the works of authors, artists, and inventors. It grants the creator exclusive rights to use, distribute, and profit from their work, typically for a specified period. But there’s a catch: copyright protection traditionally applies only to works that are original and created by a human. Courts around the world have generally held that a “human author” is a fundamental requirement for copyright.

However, with the advent of AI systems capable of generating creative content, this foundational principle is being challenged. AI tools like OpenAI’s ChatGPT, Google’s DeepDream, and image generators like DALL-E are producing content that seems to rival human creativity. These AI-generated works raise important questions: Can these works be considered “original”? And if so, can an AI ever be recognized as the creator in the eyes of the law?

Is AI-Generated Content Truly Original?

One of the cornerstones of copyright law is the concept of “originality.” Traditionally, this implies that a work must be the result of an individual’s intellectual effort and creative choices. In countries like the United States and the European Union, the law requires that a work reflect the author’s personal expression—essentially, their “personal touch.” So where does this leave AI?

Consider this: an AI-generated painting in the style of Van Gogh. Is it an original creation, or merely a replication? AI systems, after all, rely on vast amounts of data and algorithms to create their outputs. They analyze existing works, learning patterns, styles, and structures, and then use this learned knowledge to generate new content. But does this constitute true originality, or is it just a sophisticated form of mimicry?

Take, for instance, the AI-generated painting “Portrait of Edmond de Belamy” that sold for $432,500 at a Christie’s auction in 2018. The artwork was created by an algorithm developed by a Paris-based art collective. However, the algorithm was trained using a dataset of 15,000 portraits spanning several centuries. So, is the painting an original creation of the AI, a derivative work of the human creators who built the algorithm, or does it belong to the artists whose works were used in the training dataset? The ambiguity here demonstrates the complexity of attributing originality in AI-generated content.

The Human Element: Is There an Author?

Another pillar of copyright law is the requirement of an identifiable author. Traditionally, the “author” is the individual who creates the work, exercising their skill and labour to produce something unique. In the context of AI, the question becomes: who, if anyone, is the author? Is it the developer who designed the AI, the user who provided the prompts, or the AI itself?

Some might argue that the developer of the AI system should be considered the author since they created the tool that generated the content. Yet, the developer doesn’t typically have control over the specific output the AI produces in each instance. They set the rules and provide the framework, but the actual content generation happens autonomously and, to complicate things even more, the content generation is usually instructed by a human user.

If simply prompting an AI to generate content qualifies someone as an author, does that diminish the significance of the creative process that copyright aims to protect?

Accordingly, others may point to the user who prompts the AI, arguing that they guide the creative process by selecting inputs and fine-tuning the AI’s behavior. For example, a person who uses an AI tool to generate a piece of music by providing specific stylistic guidelines may be seen as playing a crucial role in the creation of the work. However, this perspective is also fraught with challenges. If simply prompting an AI to generate content qualifies someone as an author, does that diminish the significance of the creative process that copyright aims to protect?

And what about the AI itself? Can a machine be an author? Current copyright laws generally do not recognize non-human entities as authors. This has been illustrated in cases like the infamous “monkey selfie” incident, where a photograph taken by a macaque was deemed ineligible for copyright because the monkey was not a human author. Similarly, in the United States, the U.S. Copyright Office denied copyright protection to “A Recent Entrance to Paradise,” a work autonomously created by an AI. According to the office, copyright law “only protects the fruits of intellectual labor that are founded in the creative powers of the human mind.” In other words, machines, no matter how intelligent, do not qualify as creators.

The UK’s Unique Approach: Computer-Generated Works

Interestingly, the United Kingdom has taken a somewhat different approach. Under the UK’s Copyright, Designs and Patents Act 1988, there is a provision for “computer-generated works.” It states that if a work is generated by a computer in circumstances where there is no human author, the author is deemed to be the person by whom the arrangements necessary for the creation of the work are undertaken. This wording is open to interpretation but generally points to the person who set up the AI system to create the work.

While this offers a pragmatic solution, it still doesn’t fully address the complexities of authorship in the AI context. As AI systems become more autonomous, producing content with less direct human intervention, the line between the person “undertaking arrangements” and the creative output becomes increasingly blurred.

The Policy Debate: Why Does It Matter?

Why is this issue so crucial? For one, the question of copyright protection impacts how AI-generated works can be used, shared, and monetized. Without clear ownership rights, it becomes challenging to protect these works from unauthorized use. Moreover, the lack of clarity could stifle innovation. Developers, artists, and businesses may hesitate to invest in AI-driven creativity if they cannot secure intellectual property rights for the output.

On the other hand, granting AI-generated works copyright protection could have profound implications. If AI-created content can be copyrighted, who stands to benefit? Developers, users, or corporations? And what happens to the vast amount of data—often copyrighted material—that these AI systems are trained on? Should the original creators of that data be entitled to some form of compensation or recognition?

For example, when AI music-generating tools create compositions reminiscent of established artists, should those artists have a say in how their styles are used? In April 2023, an AI-generated song mimicking the voices of Drake and The Weeknd was released, raising ethical and legal questions. The song, “Heart on My Sleeve,” quickly gained popularity, but neither artist had authorized the use of their likenesses. Such incidents highlight the ethical minefield that accompanies AI-generated works.

Current Status and Future Directions

As it stands, the legal landscape around AI-generated works is still in flux. Different jurisdictions have varying interpretations, and no international consensus has been reached. The EU, for instance, has not explicitly provided for AI-generated works in its copyright laws, but the European Commission is actively studying the issue, signaling that changes could be on the horizon.

In contrast, the United States maintains a firm stance that copyright protection requires human authorship. While developers and users of AI can still protect their algorithms and software through patents and trade secrets, the content produced by these systems remains in a legal gray area.

One possible future direction is the establishment of a new category of rights specifically tailored for AI-generated works. This approach could acknowledge the unique nature of AI creativity while offering a balanced framework for protection. For instance, a “related rights” model could grant a limited set of exclusive rights to those involved in the AI’s creation and use without fully equating it to human authorship.

Another avenue could involve shared copyright, where rights are divided among developers, users, and possibly even the original data contributors whose works were used to train the AI. Such a system would require careful consideration and clear guidelines to avoid conflicts and ensure fair compensation.

So, Can Machines Own Creativity?

The question of whether AI-generated works can be protected by copyright law challenges our traditional notions of creativity and ownership. While AI systems can produce content that rivals human creativity, the lack of direct human involvement in the creative process raises profound legal and ethical questions. Current copyright frameworks, largely built around the idea of human authorship, struggle to accommodate the complexities of AI-generated works.

As AI continues to blur the lines between human and machine creativity, lawmakers, artists, and technologists will need to engage in a nuanced dialogue. They must balance the need to protect and incentivize creativity with the imperative to prevent overreach in granting rights to machines. Will we see a future where AI-generated content is afforded its own category of protection, or will copyright remain a bastion of human creativity? Only time will tell, but the conversation is far from over.

author avatar
Nicola Taljaard Lawyer
Lawyer - Associate in the competition (antitrust) department of Bowmans, a specialist African law firm with a global network. She has experience in competition and white collar crime law in several African jurisdictions, including merger control, prohibited practices, competition litigation, corporate leniency applications and asset recovery. * The views expressed by Nicola belong to her and not Bowmans, it’s affiliates or employees

This content is labeled as created by a human - more information