The Legal Dilemma of AI-Generated Copyright
Artificial Intelligence has created machines with the capacity to think, analyze data, make unaided decisions, and solve challenging tasks without human intervention. The main legal challenge today revolves around the ownership of the output generated by AI programs. It is worth noting because until recently, “copyright law excluded utilitarian works that had functions beyond the conveying of information or the displaying of some sort of appearance” (Ezema & Ibekwe, 2022, p. 110).
Most legal frameworks allow for the inclusion of software as copyrightable material. Resolving copyright issues regarding computer-generated material will necessitate interpreting court decisions regarding the utilitarian quality of software. This is primarily because copyright law pegs ownership in authorship and clearly defines the elements that constitute the ownership of a physical object or intellectual property.
AI systems are increasingly becoming important in life, given their importance in running systems and solving problems. For instance, Hu (2023) notes that ChatGPT is the world’s fastest-growing application, having had over 100 million users over two months. A variety of factors support the push for copyright ownership by AI systems.
Firstly, there is a need to hold a specific entity accountable when things go wrong (Chesterman, 2020). The need for accountability is demonstrated in the Getty Images and Stability AI case, in which the latter was sued for using an estimated 12 million copyrighted images without permission from the copyright owners (Garon, 2023). The objective is to address gaps in accountability and the inherent opacity that characterizes AI systems. Secondly, entities need to be rewarded for their contributions in specific fields. Given the overwhelming influence of AI systems on human life, the debate on intellectual property rights is, therefore, contentious and demands urgent attention.
The Debate on AI Legal Personality
The debate on AI and copyright ownership has been speculative in recent years. However, Saudi Arabia’s decision to grant citizenship to Sophia, a humanoid robot, and Tokyo’s move to offer residency to an online system believed to have the persona of a seven-year-old boy brought AI issues to the fore (Chesterman, 2020). The main issue today is the status of AI as a legal person. The current legal environment recognizes natural persons, human and juridical persons, and non-human entities for which the law has granted specific duties and rights. It must be noted, however, that juridical persons are largely an aggregation of human actors.
A growing school of thought advocates for considering AI as a legal personality to facilitate the resolution of specific liability issues. There is an argument for the trial of robot criminals who will either be reprogrammed or destroyed when found guilty of committing crimes (Chesterman, 2020). While such arguments are fundamentally beyond the scope of this paper, they highlight the significance and need for considering AI’s status in the legal framework.
Challenges of Applying Traditional Copyright Laws to AI
The context in which the argument regarding the AI copyright issue is based is characterized by poorly defined laws that create confusion in concepts and do not have clear implementation measures. It is worth noting that the authorship or ownership of computer-generated works has seldom been in doubt. The traditional belief was that because a computer is a man-made device, it is a tool created by a human, meaning that the authorship of the resultant works would be attributed to the human in question (Ezema & Ibekwe, 2022). The extensive growth of AI is challenging the notion above, even though copyright legislation necessitates the presence of a human element to consider copyright ownership. Numerous jurisdictions worldwide only recognize natural persons and other legal entities, such as corporations, as potential copyright owners.
The protection of the AI system involved in generating a specific work is a matter of concern. The current legal environment does not protect the AI system but rather the code it is based on, as outlined in the European Union’s Software Directive (Spindler, 2019). In addition, the data used to train the system is not protected, which has created a significant problem for copyright infringement law in many instances of AI training. Works produced by AI systems are also unprotected by copyright law in numerous jurisdictions worldwide. The failure to recognize AI systems as legal entities entitled to specific rights and responsibilities complicates the provision of copyright entitlement to such entities, even though many of them mimic human behaviors.
Various global leaders have reaffirmed their position that humans are necessary for considering copyright ownership. For instance, the U.S. Court of Appeals emphasized in the “Monkey Selfie” case, in which a photographer published selfies taken by a monkey, that animals have no copyright ownership rights (Ezema & Ibekwe, 2022). In addition, Stephen Thaler’s suit against the copyright office for denying an AI system authorship of artwork demonstrates the limited scope of the current legal framework (Zirpoli, 2023).
The anthropocentric perspective highlights the fact that authors must inherently be human. The U.S. Copyright Office requires human authorship for the registration of specific works. It will exclude any entries produced by machines or mere mechanical processes that exclude the input of a human author (Zurth, 2020).
The necessity of human authorship has been echoed by the European Court of Justice, the Berne Convention, Switzerland, Australia, and Japan (Zurth, 2020). The above position has been criticized by various individuals and entities that believe AI should be accorded specific rights. The prevailing argument is that human and non-human entities should be assigned authorship and ownership rights.
Numerous jurisdictions base their copyright legislative frameworks on economic rewards and numerous moral theories. The aforementioned theories seldom consider the ownership of copyright by AI as a possibility. The economic theory posits that for a novel work to be generated, the anticipated return must equal or surpass input costs (Aziz, 2023). In essence, the author of a specific work must be able to recoup the expenses incurred in generating specific works.
AI systems are unique because it is impossible to incentivize them to produce better work, making economic incentives meaningless to them. Moral rights offer credit to authors and protect their work from unnecessary alteration (Aziz, 2023). AI cannot comprehend or appreciate moral rights, which have immense value to humans. The intersection of moral rights and economic theory challenges the consideration of AI as a legal entity capable of ownership and authorship.
Another contentious issue concerns whether or not AI-generated works are personal intellectual creations. Contemporary AI systems are characteristically unpredictable, which means that the deterministic approach that governs the use of digital tools does not apply (Spindler, 2019). The use of the software is traditionally attributed to the author, given that the outcome is foreseeable. However, the unpredictability of using AI creates a significant challenge. The arguments against the copyright of AI-generated material have been bolstered by the view that human beings have lost control of the creative process (Hugenholtz & Quintais, 2021).
While various schools of thought state that AI is intelligent, it is not comparable to human will in the legal sense. In addition, the fact that AI systems are, to a certain degree, incapable of setting goals and preferences, the individual making use of the AI system is considered the author (Spindler, 2019). The AI system is, in this case, viewed as a tool rather than an independent entity.
Numerous authors in the field oppose the view mentioned above. For instance, Kriebitz and Lütge (2020) surmise that the AI algorithm should own AI creations. The algorithm is unpredictable and thus capable of independent thought and creativity. The main issue with the aforementioned position is that the current legal framework does not recognize entities that lack direct human input as copyright owners. The limitations in direct human input are seen in using AI-generative texts to facilitate the creation of specific projects or ideas.
Common tools such as Wordcraft are not considered coauthors in texts created by humans, even though the AI system was directly involved (Ippolito et al., 2022). AI systems are not recognized as legal entities in many of the world’s legal environments. Therefore, although artificial intelligence can be unpredictable, the person using it is responsible for setting the parameters that determine the specific outcome of the work and is legally considered the author.
Toward a Hybrid Legal Framework for AI Ownership
Owing to the technology’s increasing development and influence, the AI copyright challenge needs to be addressed. The following recommendations are intended to address the issue and facilitate the development of an effective and comprehensive legal framework. It is recommended that a hybrid option that considers AI as a legal entity be adopted by legal systems worldwide. Combining specific strengths while limiting the weaknesses that characterize current legal contexts is important.
A hybrid system meaningfully preserves the author’s right to a specific creation. The proposed solution provides an AI system stakeholders with a legal personality, facilitating the formation of an AI Legal Entity. The entity may include program designers, users, the AI system, and company owners.
Developing an Artificial Intelligence Legal Entity addresses key fundamental changes associated with AI copyright ownership. Firstly, the ownership of original work is addressed through various options, such as creating a unique shareholders’ agreement that serves as the governing document for the newly formed entity. Each member may be granted ownership based on their contribution to the original work. This facilitates compliance with international copyright laws that allow artificial personalities, such as corporations, to own original works. The proposal also addresses issues associated with accountability and liability in cases of legal retribution, such as copyright infringement.
The second issue that will be addressed by forming an Artificial Intelligence Legal Entity is the necessity of economic incentives and rewards. The fact that the AI’s work will be protected means that the AI Legal Entity will continue to reap the rewards of the generated material. The inclusion of human programmers, owners, and shareholders means that the collective entity will desire to produce more work of higher quality.
It is also worth noting that Artificial Intelligence Legal Entities provide room for the inclusion of artists, musicians, and writers who were directly involved in the training of the AI system. Creating a collective entity provides an avenue and a platform through which AI work can be monetized. Recommendations for further studies include the assessment of the challenges associated with considering AI as a legal entity and analyzing opportunities for abuse in AI-associated copyright provisions.
Artificial entities or personalities are concepts created by humans who also developed the idea of property. Copyright law faces challenges with regard to AI systems, including liability challenges and the allocation of benefits accrued from original works and creations. The fact that AI systems are capable of independent thought has created an argument for their consideration as legal entities that must bear responsibility and reap the benefits of their actions. Given AI’s important role in contemporary society, the current framework is marred by challenges that must be overcome. Hybrid models that create legal entities in which AI systems and their creators are considered copyright owners help address challenges associated with the ownership of AI-generated content.
Reference List
Aziz, A. (2023) ‘Artificial intelligence produced original work: a new approach to copyright protection and ownership’, European Journal of Artificial Intelligence and Machine Learning, 2(2), pp. 9–16. Web.
Chesterman, S. (2020) ‘Artificial intelligence and the limits of legal personality’, International & Comparative Law Quarterly, 69(4), pp. 819–844. Web.
Ezema, A.O. and Ibekwe, C.S. (2022) ‘Ownership of copyright in works of artificial intelligence: need for a legal framework’. Awka Journal of Public and Private Law, 10(0), pp. 108–125. Web.
Garon, J.M. (2023) ‘A practical introduction to generative AI, synthetic media, and the messages found in the latest medium’. SSRN Electronic Journal. Web.
Hu, K. (2023) ChatGPT sets record for fastest-growing user base. Reuters. Web.
Hugenholtz, P.B. and Quintais, J.P. (2021) ‘Copyright and artificial creation: does EU copyright law protect AI-assisted output?’, IIC International Review of Intellectual Property and Competition Law, 52(9), pp. 1190–1216. Web.
Ippolito, D. et al. (2022) ‘Creative writing with an AI-powered writing assistant: perspectives from professional writers’, Human Computer Interaction. Web.
Kriebitz, A. and Lütge, C. (2020) ‘Artificial intelligence and human rights: a business ethical assessment’, Business and Human Rights Journal, 5(1), pp. 84–104. Web.
Spindler, G. (2019) ‘Copyright law and artificial intelligence’, IIC International Review of Intellectual Property and Competition Law, 50(9), pp. 1049–1051. Web.
Zirpoli, C.T. (2023) ‘Generative artificial intelligence and copyright law’. Web.
Zurth, P. (2020) ‘Artificial creativity? A case against copyright protection for AI-generated works’. UCLA Journal of Law and Technology, 25(2), pp. 1–22. Web.