This discourse centres on the interesting collision of artificial intelligence in the world of intellectual property law and protection; and it beacons on “originality and novelty tests” of the right. Schalkoff agreeably refers to artificial intelligence (AI) as “a field of study that seeks to explain and emulate intelligent behaviour in terms of computational processes” (Schalkoff, 1990). It a form of technology and could hence, analyze data, think, speak, recognize, make independent decisions, solve complex problems, learn, even feel and react without any human help (Krzisnik, 2019). Generative AI is hence, a form of artificial intelligence that can act on the foregoing strengths to create new content. Put quite technically, Generative AI refers to “a category of ML algorithms that generate new outputs based upon the patterns within data on which they have been trained” (Brown, 2021) Intellectual Property Rights (IPR) protection is crucial for the promotion and encouragement of innovation and creativity as ways of rewarding effort, intellectual labour, and skill (section 1(a), Copyright Act 2022). Hence, a book, sculpture, music, technology, distinctive marks, distinct methods of a business (section 2 Copyright Act 2022), can earn the protection of the law to only allow certain rights (like; use, transfer, etc) to the original creator of such work under certain IPRs (like copyright, patent, Trademark, Trade Secret as the case maybe). Further, this ensures that any other person who does an act capable of varying the protections accorded the original creator, would either criminally or in civil law, be culpable depending on the circumstances of such breach.
The topic of intellectual property (IP) law has taken on new dimensions as a result of the advent and development of AI. AI technology has progressively become crucial to several Nigerian economic sectors, including manufacturing, healthcare, and banking, as the country’s economy continues to expand. As a result, it is essential that Nigerian IP regulations stay current with AI technology advancements and foresee potential obstacles in the IP space. The central query here is whether Nigerian law ought to provide IP protection for AI-generated works.
Nigerian copyright law, for example, considers the author of a work to be the individual who created it, and AI-generated works have no author in the traditional sense as they are a product of computational processes external to an individual’s creativity. When it comes to AI-generated works, the question of originality (and in perfect circumstances; novelty) takes the spotlight. To what extent can the denial be? Can there be exceptional instances? Is there a strong legal regime? What needs to be done in law and practice? These and more, are the questions this piece intends to address through a doctrinal methodology. Taking on these, would show that the growth of AI technology has brought about new challenges that Nigerian IP laws must address.
This article would be of immense interest to intellectual property (IP) students, scholars, law-makers, IPR owners, and government bodies
2.0 Synopsis of IP Framework under Nigerian Laws
What is the current framework for IPR protection in Nigeria? And what does these provisions portend in the Age of Generative AI? These are questions worthy of prudent enquiry in this section. In Nigeria, the framework for intellectual property rights (IPR) protection is based on a number of statutory regulation sourced from a number of legislation. These includes; the Copyright Act 2022 which was enacted on April 6, 2022, and signed into law by the president on March 17, 2023. The Copyright Act 2022, makes provisions for the regulation, protection, and administration of copyright, replaces the Copyright Act, Cap C28, Laws of the Federation of Nigeria (LFN), 2004. Other major laws are; Patent and Designs Act, 1988 Cap P2 LFN 2004 and the Trademarks Act, Cap T3, LFN 2004.
The Copyright Act 2022, for instance, which establishes the requirements for copyright protection. In s2 of the Copyright Act 2022, the following works shall be eligible for copyright: (a) literary works; (b) musical works; (c) artistic works; (d) audiovisual works; (e) sound recordings; and (f) broadcasts. A literary, musical or artistic work in Nigeria is only protected by copyright under section 2(a) & (b) of the Copyright Act 2022 if it satisfies a number of requirements, including originality, fixation to a recognized and specific medium of expression, and sufficient effort to establish the work’s original status (Stren and Blan, 2023). This is thus because ideas are not protected by copyright; rather, how ideas are conveyed is. For example, considering that the ChatGPT generated materials lack originality and were not developed by a person who intended to or represented claims over the output, it is unlikely to be protected by copyright in Nigeria. The needed regulatory attention is urgent in this area since, this is gravely permeating the schools and workplaces as the “easy-peasy” methods of producing academic papers to fulfill course requirements, routine work-based tasks, data-entry, content creation, etc as the case may be.
Likewise, in accordance with section 1(1) of the Patent & Designs Act, an invention must be novel, the product of an inventive step, and suitable for industrial use in order to qualify as patentable. A trademark is also defined as “A word, letter, label, numeral, color, signature, device, or any combinations of words, letters, labels, or signatures that identify and distinguish the source of the goods or services of one manufacturer from those of others in the course of trade” in section 67 of the Trademarks Act. These form the basis of the creation, protection, and regulation of these rights and more connected thereto. The 3 forms of IP noted in this section are not all there is, rather a way to achieve brevity. At all times, whatever would be protected in Nigeria as IPR, must have been shown to be in conformity with the provisions of these laws as well as Directives from the Commissions established to champion their enforcement.
3.0 Legal Concerns of IPR in Generative AI Era
The implementation of patent law is another problem with AI and IP rights under Nigerian law. Patents are issued in Nigeria for innovations that are new, inventive, and have industrial use. Due to their “non-human” creator character, AI-generated innovations like algorithms that produce “computer software” may not fall under the standard parameters of patent law. Along with “novelty,” determining the “inventorship/developer/originator” of an invention is another patent-related concern in Generative AI that constitutes one of the stubborn factors in AI-generated innovations. Gleaned from the ruling of the United States Federal Circuit Court (Horton & Kim, 2022), that inventors must be natural humans and exclude artificial intelligence technology, is adopted. While this stands, can the person to whom the algorithm is generated for, be clothed with patent rights? While this is a white area, consideration must be had on the novelty of such generated output, since the possibility of the AI generating the pathways and features of existing software is not ruled out. Further research or law and policy-making in Nigeria should consider questions about the extent to which AI technology can be credited with invention, and how this impacts IP law.
It becomes essential to notice in points 3.1 and 3.2, above, that Nigerian law has not yet addressed the problem of ownership of data produced by AI. AI-generated data is typically not protected by intellectual property laws since it lacks the conventional originality or novelty necessary for copyright or patent law protection.
It is obtainable that users might unintentionally publish content that is too similar to a copyrighted work because they are unaware of all the works that are protected by copyright (Khan & Wegrzyn, 2023). The user’s reputation may suffer substantial harm as a result, and there may be legal liabilities. One of the most significant hurdles to IPR protection in the era of Generative AI is the possibility for copyright infringement in AI output and the usage of IP-protected resources without acknowledgement or license. These issues will need constant regulatory attention as AI is evolving and is getting more incorporated into society’s daily activities and innovation processes. This s particularly pertinent given that generative AI systems may produce work that is in violation of copyright laws and exposes the individual/organization that distributes or publishes it to charges of copyright infringement. In this regard, the viability of a licensing and/or acknowledgment/attribution system (depending on the right(s) in question) is recommended as a viable alternative to copyright infringement and needs to be explored. Existing exceptions for text and data mining from protected sources and with respect to infringement also need to be examined.
A number of cases have identified intellectual property violations in the domain of AI. For instance, in the case of Andersen v. Stability AI Ltd (3:23-cv-00201-WHO (N.D. Cal.). three artists brought legal action against AI art generators for using their creations without their permission or payment in a data training set. Plaintiffs claim that defendants downloaded copyrighted works from the internet and used those works to train their AI image generation models without obtaining a license, constituting direct copyright infringement. They also claim that by doing so, they violated the Digital Millennium Copyright Act (DMCA) by removing copyright management information (such as copyright notices, titles, and author information). See also; Planner 5D v. Facebook, 3:20-cv-08261, Dkt. 1 (N.D. Cal. Nov. 23, 2020).
An unsettling issue has been raised about the potential to use AI systems to produce unlicensed copies of goods or services that have been trademarked (Lari-Williama, 2023).
4.0 Any Exceptions?
However, AI-generated works could at times still be regarded as unique and deserving of IP protection. For instance, the work produced by an AI that was trained to create something new and original by its programmers may be considered satisfactory of the statutory requirements of “originality and novelty”, and the programmer may be regarded as the creator with consequential rights. As a result, such works are nonetheless subject to protection under Nigerian copyright law, and their authors may be given rights. To found the possibility of an exception, Lari-Williams questions if the applicability of fair use/fair dealing principles, which are used by AI researchers, startups, and big tech corporations, is obvious in the context of Nigerian legislation given that AI relies on pre-existing works (supra). Fair use/fair dealing as one of the IPR infringement defences, apply to excuse a defendant from infringement liabilities of certain unauthorized IP use. Whether this should be held applicable with respected to unattributed contents generated by AI; is a matter worth considering by the Law-Makers and Regulators. Whichever they finally hold, it can only excuse the Defendant in infringement actions, not qualify the content for IPR protection.
5.0 Conclusion & Recommendations
Protection of intellectual property rights (IPR) is essential for encouraging and promoting creativity and innovation. The current Nigerian system for IPR protection, however, faces considerable deficiencies with the introduction of generative AI. In order to address the issues presented by generative AI, the first section of the study included a general summary of Nigeria’s present IPR protection system. The issues posed by generative AI and how they are impacting intellectual property protection were then explored. In order to secure proper protection in the era of Generative AI, the study concluded by suggesting potential updates or revisions to the current IPR rules.
As a result, this article adds to the ongoing discussion about IPR protection and generative AI and offers useful information to academics, stakeholders, and policymakers. Therefore, in order to inform the establishment of future IP rules in Nigeria, the government must take into account the ramifications of AI-generated data and patent ownership. Overall, a thorough grasp of AI growth and its effects on Nigerian IP law is required to guarantee that inventors and creators have the necessary legal protection to profit from their labor and inventiveness. In order to solve these issues associated to generative AI, it is necessary to amend or update intellectual property laws through legislative actions. However, given that Nigeria’s Copyright Act was just recently brought into law in 2022, it is anticipated that legislative actions surrounding copyright would be difficult to implement in the near future. This does not, however, discredit the suggested legislative measures.
Further, the creation of a dedicated AI division inside the Nigerian Copyright Commission (NCC) and other IP Government institutions; is advised in order to assist inventors in navigating the IP system in light of the realities of AI. This is a feasible option pending the enactment or update of IP legislation that takes this into account. It will also aid national, cross-national cooperation and exchange of practice, procedures and technology.
Again, workplaces and educational facilities should be weary of the impacts of Generative AI in producing academic papers and materials. Plagiarized manuscripts produced as academic papers are not to be considered scholarly and should not be encouraged. Likewise, workplaces should exercise caution in knowing what they create and what is drawn from other sources to avoid liability. Hence, individuals, workplaces, and educational facilities should imbibe acute consciousness of IPR while embracing technology use.
Finally, tech companies are applauded for their innovations through AI and it is recommended that they make improvements to ensure that generative contents are attributed to the sources they are generated from and licenses (where applicable) obtained for use of any material or knowledge. This way, the attendant issues identified in this article can at least, be half-solved.
Author: Joshua Olewu
DELTA Data Protection & Compliance Academy & Consulting – firstname.lastname@example.org