top of page
Search

The Unresolved Dilemma: Artificial Intelligence and Copyright Ownership Dispute

sonalimukhia2002
Aug 21
8 min read

Author: M. N. Varsha, Presidency University


Introduction

The rapid development of artificial intelligence has posed serious challenges to the traditional legal frameworks, especially in the area of intellectual property. Copyright law, which has always been based on human authorship, has to deal with a new question: who gets the copyright for a work produced by an AI tool? The urgency of this question stems from the fact that AI technology has evolved from being used as a mere tool to becoming a self-sufficient creator of literary and artistic works which would otherwise be produced with human intellectual input. The law does not seem to cope with the new realities, as different jurisdictions adopt different strategies which lead to ambiguity for creators, investors, users, etc. The aim of the article is to analyse the legal status of AI-generated works in light of recent American judicial rulings as well as strategic moves regarding AI copyright laws in Europe. The author points out that while existing law is conceptually sound, it is not going to work in practice. Finally, he calls for a differentiated approach to the classification of AI-generated works depending on the degree of human creativity involved.


The Foundations of Copyright and Human Authorship

Copyright protection in general presupposes the presence of human creativity. According to the Constitution of the United States, the copyright of Congress presupposes the possibility of protecting the rights of artists and inventors and thus reflects the link between copyright and the role of humans. This idea has played a role in all the processes of statutory interpretation and decision-making regarding copyright. For example, the definition of ‘works of authorship’ in the Copyright Act of 1976 is also based on the definition which implies the existence of human creativity during the process, including literary works, musical works, and pictorial, graphic, and sculptural works. The traditions of the U.S. Copyright Office show the same tendency as it is necessary to show ‘the minimal level of creativity’ of the works created by a person.

This approach is based on the common law principle. As it was defined in the case of Feist Publications, Inc. v. Rural Telephone Service Co. by the Supreme Court, “the sine qua non of copyright is originality,” which means that the work has to have at least a minimal degree of creativity. Despite the fact that this standard is rather low, it means that there is a need for a human being as a source of such creative decisions. The European Union’s attitude towards intellectual property does not differ much from the one of the American Supreme Court, because its Court of Justice holds the same opinion. All the standards concern the capacity of works to meet the requirement of “author’s own intellectual creation,” which has been explained by the courts in terms of human activity. The above frameworks have shown their applicability to various technological advancements; however, the core of the process is still the same – it is human intellectual labour.


The United States Position: Refusal and Resistance

The U. S. Copyright Office has consistently denied registration of AI-generated works. In the case of A Recent Entrance to Paradise , the Copyright Office’s Review Board rejected a Creativity Machine’s claim due to the lack of human input into the creation of the work. The Review Board held that copyright law “requires a connection between the human mind and the creation,” and therefore works with no “human creative input” are not copyrightable. Similarly, the Copyright Office granted copyright for the way in which the images were assembled in the comic Zarya of the Dawn , but denied copyright for the images themselves, which were generated by the Midjourney AI.

The courts have largely sided with the Copyright Office. In the case of Thaler v. Perlmutter , the U.S. District Court for the District of Columbia ruled in favour of the Copyright Office’s refusal to register an AI-generated work. The court held that “the notion of authorship is necessarily a human concept,” reasoning that copyright law’s “focus on authorship and the rights of authors” and long history of involving humans in the creative process meant that AI-assisted works could not be copyrighted. According to Judge Beryl A. Howell, the copyright law’s “language, structure, and overall scheme, along with the consistent implementation and interpretation by the Copyright Office, make it clear that human authorship is a prerequisite for copyright protection.”

This outcome has significant implications for AI-generated works. With very few works made by AI with no human input receiving copyright, those without copyright will enter the public domain. This may drive developers to find other ways to protect their works, such as through technological devices or contract law. At the same time, however, the ruling upholds copyright law’s basic tenets.


The United Kingdom and European Approaches: Variations on a Theme

The UK has adopted an innovative solution in terms of the copyright law system. According to section 9(3) of the Copyright, Designs and Patents Act of 1988, “in the case of computer-generated works, the author shall be the person by whom the arrangements necessary for the creation of the work are undertaken.” Although this regulation is formulated way before the contemporary systems of artificial intelligence were invented, it appears relevant in the present-day situation. The statutory formula gives courts an opportunity to assign copyright to those people who have some control over AI systems. Nevertheless, it has rarely been interpreted by courts in the context of modern artificial intelligence.

The EU method is based on the concept of “author’s own intellectual creation,” which was explained in the case Infopaq International A/S v. Danske Dagblades Forening. The Court of Justice of the European Union ruled that copyright can cover the work, which is original from the point of view of embodying the personal touch of the author through “free and creative choices.” Although such a criterion does not exclude AI-generated works directly, it poses problems for those works in which such choices are made by algorithms rather than people. At the same time, there is no special provision concerning the computer-generated works in the European Copyright framework yet.


The AI-Assisted Works Distinction

An important demarcation between fully AI-generated material and AI-assisted material has come up. The U.S. Copyright Office has made it clear that material wherein the involvement of AI is present along with the presence of a significant amount of human creative input can be copyrighted. In its instructions, the Office stresses the point that “the human author must have exercised control over the expression” and that “prompting a generative AI system does not create the necessary level of creative control.” But at the same time, the Office has conceded that material wherein significant human input has been made in the form of curation, arrangement, or modification of AI generated content can be copyrighted as much as human creative input has been made.

However, the implementation of the distinction mentioned above is difficult. It is not entirely clear where the border between sufficient and insufficient human participation lies, which causes doubts for creators and for those who use their creations. As was mentioned by the Copyright Office, each case will be considered individually taking into account some details like whether the ultimate creative control was performed by the human author or AI was used just as an assistive tool.


International Developments and Emerging Consensus

This issue is becoming increasingly popular at both global and national levels. The World Intellectual Property Organization has begun discussing intellectual property rights for AI and encountered inconsistencies within its member states. For instance, while many developing nations are advocating for the maintenance of the principle of human authorship to protect cultural expressions and human creativity, developed countries are promoting sui generis solutions to protect AI productions considering economic advantages and complexity of the issues around inclusion of AI productions in copyright frameworks.

The differences in the approaches taken by developed and developing countries demonstrate that there are several ways to approach the issue. For example, China is taking steps to formulate a unique approach to address the complications involved in AI-made productions. In one of the cases in 2020, Beijing Internet Court recognized the authorship of AI production with limited human involvement. The court stated that protection of copyright is possible for such productions due to possibility of identifying human input. Japanese scholars have proposed another approach to address similar issues. According to them, copyright protection for works made with AI is possible if originality of such works is determined despite questions surrounding authorship.


Critical Evaluation and Future Directions

While the present legislation might seem logical, in practice, it is rather difficult to implement. It will not only be problematic but also will negatively affect the further development of AI technologies because leaving their works without any copyright protection implies that there will be no sense in making efforts to develop new systems capable of creating amazing works of art because they will immediately become publicly accessible. However, copyrighting AI-made products will contradict the very nature of the copyright concept as such, as its aim is to protect human creations.

It could be suggested as a possible way out of the situation to provide some form of sui-generis protection for AI-made products. It could include shorter copyright duration or some other specific limitations to the way the copyrights can be used or some other options. The sui-generis protection exists for the EU databases as a result of the Database Directive according to which database creators can benefit from exclusive rights even if they do not meet the criteria of the authors.

The further suggestion is to reformulate the existing regulatory framework governing the copyright protection of AI-generated works and extend the protection that is provided under said framework. The existing practices of courts and administrative bodies should provide clear guidelines to those interested in the laws regulating copyrights, especially since the legal frameworks should remain the same. While the Office of Copyright has already made some moves in this direction, additional clarity should be provided in the context of innovation processes that are related to the use of AI technologies.

Moreover, the market is a good way to solve the issues observed in the sphere of copyrights and alternative tools can be used to reach the desired results. The use of contract law allows defining ownership rights over AI-generated works at times when the creators of technology are in a strong position and can formulate the contract terms. Trade secret law can also provide some protection since its mechanisms are not so strict as in the case presented above yet the trade secret law itself requires taking reasonable measures to protect secrets.


Conclusion

One of the most significant legal issues in the contemporary digital world is the merging of artificial intelligence with the issue of copyright. Despite being faced with new technologies, the traditional view based on the need for human creativity has remained strong. Courts and other legislative organizations have upheld the requirement that only humans can be recognized as authors of works and consequently given the right for copyright protection.

The clear distinction between human-created and AI-generated products can serve as a starting point for a more intricate system of regulations that will preserve the meaning of copyright protection. Having in mind that artificial intelligence is developing every day, it seems reasonable to create new laws that will allow to combine the necessity of protecting human creativity and protecting the legitimate interests of economic entities making profit from AI products.


References

Cases

  • Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340 (1991).

  • Infopaq Int’l A/S v. Danske Dagblades Forening, Case C-5/08, ECLI:EU:C:2009:465 (C.J.E.U. July 16, 2009).

  • In re A Recent Entrance to Paradise, U.S. Copyright Off. Rev. Bd. (Feb. 14, 2022).

  • In re Zarya of the Dawn, U.S. Copyright Off. Rev. Bd. (Feb. 21, 2023).

  • Thaler v. Perlmutter, 687 F. Supp. 3d 140 (D.D.C. 2023).


Statutes

  • 17 U.S.C. §§ 101-810 (2018).

  • Copyright, Designs and Patents Act 1988, c. 48, § 9(3) (UK).


Books and Journal Articles

  • Jane C. Ginsburg, People Not Machines: Authorship and What It Means in the Berne Convention, 44 Colum. J.L. & Arts 395 (2021).

  • James Grimmelmann, Copyright for Literate Robots, 58 Ariz. L. Rev. 657 (2016).

  • Bernt P. Hugenholtz & João Pedro Quintais, Copyright and Artificial Creation: Does EU Copyright Law Protect AI-Assisted Output?, 50 Int’l Rev. Intell. Prop. & Competition L. 279 (2019).

  • Christina Mulligan & Timothy B. Lee, Artificial Intelligence and the Future of Copyright, 108 Calif. L. Rev. 1479 (2020).

  • Pamela Samuelson, Allocating Ownership Rights in Computer-Generated Works, 47 U. Pitt. L. Rev. 1185 (1986).


Reports and Official Documents

  • U.S. Copyright Office, Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190 (Mar. 16, 2023).

  • World Intell. Prop. Org., WIPO Conversation on Intellectual Property and Artificial Intelligence: Revised Issues Paper, WIPO/IP/AI/2/GE/20/1 REV. (2020).


Online Sources






Comments

bottom of page