Artificial Intelligence and The Conflict with Copyright Law

Artificial Intelligence and The Conflict with Copyright Law

Written by Naina Das [Institute of Law, Nirma University]

Introduction

Artificial intelligence (AI) is rapidly transforming the way we create, distribute, and consume content. With the ability to generate, analyse, and interpret vast amounts of data, AI has enabled us to automate many aspects of the creative process. However, this powerful technology has also raised a host of legal and ethical issues, particularly in the context of copyright law. As AI continues to advance and become more widespread, it is becoming increasingly clear that existing copyright laws are struggling to keep pace with these developments. This article will explore the ways in which AI is creating conflicts with copyright law and the challenges that this poses for creators, rights holders, and society as a whole. This article will examine some of the key legal issues that arise when AI-generated works are created, shared, and used, and discuss the implications of these developments for the future of copyright law.

What is A.I and how it functions?

A.I is a self-learning technology, it is a computation technique use to design such programmes and algorithms which can perform tasks that normally require human intelligence such as visual perception, speech recognition, decision-making, and language translation. The idea is to create such a technology that can mimic human behaviour without explicit programming and human intervention.

A.I is powered by machine learning, which is the self-learning algorithm of the A.I. It functions on three elements, namely, knowledge representation, automated reasoning and natural language programming. Knowledge representation is that element which enables the machine to think by creating a structured model of information. Natural language processing is used to enable AI systems to understand human language. Both these elements help in the data collection part of the programming. Automated reasoning is that element of A.I which enables it the machine to learn from the data and apply its own reasoning for producing the outcome.[1]

Also Read: The Future of Law in the Age of AI: Navigating the Opportunities and Challenges Ahead

Another important

Another important element of A.I which allows it to function like a human is Artificial Neural Network (ANN). ANN is modelled on the structure of the human brain which enables the A.I to recognize patterns and learn from experience and development.

Intersection of A.I and Copyright Law

Copyright law provides protection to the original expression of creative works, such as literary, musical and artistic work. It grants exclusive rights to creators of such works, allowing them to control how their work is used and distributed. The intersection of AI and copyright law arises when AI is used to create or assist in the creation of creative works. However, this raises several legal issues related to copyright ownership, infringement, and fair use.

One of the main challenges in AI-generated works is determining who owns the copyright. Traditionally, the author of the work is the owner of the work and is granted exclusive rights.[2] But in the case of A.I generated work, determining who the author is a challenge, will it be the programmer, user or A.I itself. Another challenge is the potential for copyright infringement by AI systems. For producing the outcome, A.I useS a large source of data, and some of such data can be copyright protected, even if permission is taken from the owner for use of such data, there is a possibility of the outcome being substantially similar to a copyright-protected work. This raises questions about the legality of AI-generated works that may infringe on existing copyright.

Furthermore, in case of A.I generated work, the application of fair use doctrine becomes more complex. Fair use is a legal exception that allows for the use of copyrighted material without permission for certain purposes. However, it can be challenging to determine whether AI-generated works qualify as fair use. For example, if an AI system is trained to create a new work by sampling and remixing copyrighted works, it may be difficult to determine whether this use is transformative enough to qualify as fair use.

A.I as Copyright Owner

The copyright law grants protection to any original work, the originality here refers to the fact that the work should have come from the author itself. Now, whether any non-human can be regarded as author is the question here. An interesting case in this aspect was the Naruto “monkey selfie case” of 2011[3], herein issue arose over whether or not an animal can be recognized as an author. The U.S Court of Appeals held that an animal doesn’t have any standing in the court and hence cannot sue for copyright infringement. The court noted that corporations can own copyright and sue under the Copyright Act as they are recognized as legal persons by the courts and unlike animals, these entities are formed and owned by humans.

So, for a non-human to have a right as author of copyrighted work, it must be recognized as a legal entity in the eyes of law. As merely granting copyright protection will not be fruitful if the right holder doesn’t have any standing in court in order to exercise those rights.

The time to grant A.I ownership over the work generated by them might be around the corner as there are platforms that can create original work, be it related to music or paintings. Recently, a portrait of ‘Edmond de Belamy’ was created by[4] an A.I and sold at an auction for $432,000. So, a consumer platform for A.I generated art is emerging. The issues related to who shall benefit from profits of such art is a concern as it is not ethical to let the programmer or user to benefit entirely from this if the work is the autonomous creation of the A.I. Such concerns might be solved if A.I was to be granted a status of a legal person.

Can A.I be treated as a legal person? –

The author Visa Kurki, in his book ‘The Theory of Legal Personhood’[5] has discussed about the context of A.I as a legal person. According to Kurki, legal personhood refers to the recognition of a legal entity as a subject of rights and duties in the same way that human beings are recognized as subjects of rights and duties. In the context of AI, Kurki argues that granting legal personhood to AI systems would involve recognizing them as having certain legal rights and responsibilities, such as the ability to own property and enter into contracts, and the responsibility to pay taxes and be held liable for any damages or injuries caused by their actions.

Moreover, Kurki has also recognized the limitations to granting A.I legal personhood, A.I lacks consciousness and intentionality, which are key characteristics that underlie the concept of legal personhood. It may not be appropriate to grant legal rights and responsibilities to entities that are not capable of making moral judgments or acting with intentionality.

Further, granting such legal personhood to A.I would come with a bundle of issues, such as if A.I were to be given the right to own property and sue for infringement, it would still require a representative to be able to plead its case. What if A.I was the one being sued for infringement, in such case question arises as to who would pay the damages imposed.  These problems may be sort by appointing a representative to A.I but then question arises to who will be the representative, will it be the programmer or user. One way to overcome this issue is the theory of composite ownership liability. Under this concept, multiple parties, including the AI system, the company that owns the AI system, and the human beings involved in the development and use of the AI system, may all be held liable for any copyright infringement that arises from the use of AI-generated works.

Copyright Infringement by A.I and Doctrine of Fair Use

A.I uses a large source of data in generating one work, some data may be in public domain, some may be copyright protected, hence using of such data to create a work can lead to a case of infringement of copyright. It is not possible to ensure that machine is not using copyright data to learn and develop as there is no algorithm that can trace from where the machine actually learns from. However, one way to avoid this issue can be by taking permission for usage of such copyrighted data, but there is still possibility of the outcome being a work which infringes upon copyright claims. One defence that is available to such infringement is the doctrine of fair use.

Fair use allows usage of copyrighted work without permission from the owner for specific purposes such as research, teaching, news etc. The determination of fair use is based on four factors: nature of work, purpose of work, amount and substantiality of the portion used, and the effect of subsequent work on the economic value of copyrighted work.

The application of fair use to A.I generated work require careful consideration of the above mentioned four factors.

• Nature of copyrighted work – This factor looks at whether the original copyrighted work is creative or factual in nature. Generally, the more creative the work, the stronger the copyright protection. For example, an AI-generated work that uses factual data to create a new visualization may be more likely to qualify for fair use than an AI-generated work that closely copies a creative work such as a novel or a film.
• Purpose of subsequent work – This factor looks at whether the use of the AI-generated work is transformative, meaning that it adds something new and provides a different meaning or message from the original copyrighted work. For example, an AI system that uses an existing photograph to create a new artwork that is significantly different in appearance and meaning could be considered transformative.
• Amount and substantiality of the portion used – This factor looks at how much of the original copyrighted work was used in the creation of the AI-generated work. Generally, the less of the original work used, the more likely the use will be considered fair. For example, an AI-generated work that uses only a small portion of an existing work may be more likely to qualify for fair use than an AI-generated work that uses a large portion of the original work.
• Effect of subsequent work on economic value of copyrighted work – This factor looks at whether the use of the AI-generated work will harm the market for the original copyrighted work. If the AI-generated work is likely to reduce the market value or potential market for the original copyrighted work, it may be less likely to qualify for fair use.

However, even when considering these four factors, challenges may arise in determining the application of fair use as A.I generated work differs from human made creations. One limitation is that it may be difficult to determine the purpose of the work in context of A.I generated work. If an AI system generates a work that is similar to an existing copyrighted work, it may be difficult to determine whether the purpose of the use is transformative or merely derivative. Further, as it is difficult to trace which data A.I used for producing its outcome, this creates another challenge in determining whether the portion of copyrighted work used amounts to substantial use or not. Lastly, it will be challenging to determine whether the use of the AI-generated work will have a negative impact on the market for the original copyrighted work as the consumer base for A.I generated painting may be different than the consumer base of owner’s painting. Hence, the application of doctrine of fair use in context of A.I is complicated.

Suggestions

The complicity of application of fair use doctrine to A.I generated work can be reduced by making changes to the doctrine. The doctrine could be updated to better reflect the transformative nature of AI-generated works. This could include clarifying the definition of transformative use and providing guidelines for determining when AI-generated works are considered transformative.

Moreover, new categories of fair use could be created specifically for AI-generated works. For example, fair use could be expanded to include AI-generated works used for scientific or educational purposes. And the scope of fair use could be limited for AI-generated works that are used for commercial purposes. This could help to protect the market for original copyrighted works and prevent AI-generated works from undercutting the value of the original work.

Lastly, a liability framework should be developed to address all these issues. Like composite ownership liability could be imposed, or liability could be imposed by appointing representatives to A.I or it can be affixed on the programmer or the person who trained the system with copyrighted material.

Conclusion

The rapid development of A.I has revolutionized the creative process However, this development has posed a significant challenge to copyright law raising issues related to ownership, infringement, and fair use. Determining who owns the copyright of an AI-generated work, the potential for copyright infringement, and the complex application of fair use doctrine are among the key legal issues that need to be addressed. The idea of granting AI legal personhood in order to own copyrights is a potential solution that has been discussed, but it also presents its own set of challenges as it would have significant implications for future of copyright law. The intersection of artificial intelligence and copyright law is a complex and challenging issue that raises many legal and ethical questions. As AI technology continues to advance, it is clear that copyright law will need to evolve to keep pace with these developments, and it is important that society addresses these legal and ethical issues proactively to ensure that the benefits of AI are maximized while minimizing its potential negative impact on creators, rights holders, and society as a whole.

---

[1] What is Machine Learning? | IBM

[2] https://heinonline.org

[3] Naruto “monkey selfie case

[4] Edmond de Belamy’ was created by an A.I and sold at an auction for $432,000

[5] The Theory of Legal Personhood’