By Lucy Naa Asheley AFUTU
Imagine a young Ghanaian software developer. Let’s call him Kwame. Kwame just launched a groundbreaking AI tool that generates unique pieces of digital artwork. His AI system, after being trained on a vast dataset of traditional Ghanaian art and global trends, produces intricate designs that garner widespread attention. Kwame begins to sell these artworks online, receiving international recognition for the AI-generated masterpieces.
However, as the tool becomes more popular, Kwame encounters an unexpected challenge; who actually owns the rights to the AI-created artwork? Is it Kwame, the AI system, or the original artists whose work influenced the AI’s designs? Assuming Kwame gets sued by a local artist. What Ghanaian laws govern this novel technology?
This dilemma brings into sharp focus the complex questions surrounding Artificial Intelligence and Intellectual Property. As AI continues to reshape the creative landscape, Ghana’s legal framework must evolve to address these challenges. Kwame’s story is just one example of how the convergence of AI and IP law will shape the future of innovation and creativity.
Artificial intelligence and intellectual property: a converging landscape
AI represents a transformative force, replicating human intelligence in machines. This emulation enables AI systems to execute tasks that typically require human cognitive abilities, such as learning from data, discerning intricate patterns, making informed decisions, and progressively refining performance through iterative learning processes. AI systems, particularly those utilizing deep learning algorithms, can analyze vast amounts of medical images, such as X-rays, MRIs, or CT scans, to identify early signs of diseases like cancer.
The AI domain encompasses a spectrum of techniques, including machine learning, deep learning, natural language processing, and computer vision. These specialized systems are designed to execute specific tasks proficiently within predefined limitations. Examples include speech recognition, recommendation systems, and image classification.
The distinction between Generative AI and other applications of AI lies in their functionality and goals. Generative AI is designed to create new content. It uses deep learning models, like GPT to generate original text, images, music, or even code based on patterns it has learned from large datasets.
For example, a Generative AI can write a poem, compose music, or create digital art from a set of instructions. The primary goal of Generative AI is to produce new, often creative, outputs that did not previously exist while other applications of AI typically focus on analyzing data, identifying patterns, or automating tasks without generating new content.
Examples include diagnostic AI systems that analyze medical images for disease detection (e.g, identifying tumors in an MRI). General or Strong AI, aims to replicate human-like intelligence across a broad spectrum of tasks, mirroring the cognitive versatility found in human beings.
Artificial Intelligence is increasingly being utilized in the legal field to streamline processes, enhance research capabilities and improve decision making. Legal Professionals use AI-powered tools for tasks such as legal research, contract analysis, predictive analytics and document review. Artificial Intelligence Systems help identify patterns in large volumes of legal data, leading to more efficient case management and better outcomes for clients.
Intellectual Property refers to creations of the mind, such as inventions, literary and artistic works, designs, symbol, names and images used in commerce. Intellectual Property is protected by laws which covers Patents, Copyrights and Trademark, which enable people to earn recognition or benefit financially from their creations.
There are various categories of Intellectual Property which includes Patents, Copyright, Trademarks, Industrial Designs, Trade Secrets and so on. A Patent provides the patent owner with the right to decide how -or whether – the invention can be used by others. In exchange for this right, the patent owner makes technical information about the invention publicly available in the published patent document. This is governed by the Patent Act, 2003.
Copyright is a legal term used to describe the rights that creators have over their literary and artistic works. Works covered by copyright range from books, music, paintings, sculpture and films, to computer programs, databases, advertisements, maps and technical drawings.
This is governed by the Copyright Act, 2005. A trademark is a sign capable of distinguishing the goods or services of one enterprise from those of other enterprises. This is regulated by the Trademarks Act, 2004 (as amended by Trademarks Amendment Act, 2014). The pivotal question that necessitates answers pertains to how Intellectual Property intersects with Artificial Intelligence.
The intersection of Intellectual Property with Artificial Intelligence presents a unique challenge that demands thorough examination. Intellectual property encompasses a broad spectrum of intangible assets, including patents, copyrights, trademarks, and trade secrets. In the context of AI, these assets often manifest in the form of algorithms, data sets, software code, each combined to create content without direct human involvement, raising questions about who should be considered the creator.
Intellectual property and artificial intelligence’s advancement in Ghana
The current landscape of Artificial Intelligence in Ghana is gradually evolving. There is a growing interest in Artificial Intelligence technology across various sectors such as healthcare, agriculture, finance, education, law and so on.
The Ghanaian government has recognized the potential of Artificial Intelligence and digital technologies in driving economic growth and development. In 2019, the Government launched the National Digital Commerce and Payments Policy to promote the use of digital payments.
This policy has led to the emergence of various Artificial Intelligence powered payment solutions. The Honourable Minister for Finance, Dr. Mohammed Amin Adam, on March 18, 2024 also indicated Ghana’s commitment to working closely with other African countries and development partners across the world to fund Artificial Intelligence (AI) projects.
He stated that Ghana, and several other African countries are looking to collaborate with development partners such as the IMF, World Bank and the African Development Bank (AfDB) among others, to provide funding, technical and implementation support for AI projects as well as the development of national AI strategies tailored to Ghana’s unique needs.
Ghana has a growing technological ecosystem with numerous startups and companies incorporating Artificial Intelligence into their products and services. Companies like mPharma are using Artificial Intelligence to optimize supply chains in healthcare. mPharma developed an Artificial Intelligence powered inventory management system for Pharmacies, which uses predictive analytics to forecast drug demand and optimize inventory levels. This allows pharmacy owners to self-serve accurate and updated pharmacy data and generate cross-cutting sales while startups like Farmerline are using Artificial Intelligence to provide farmers with better access to markets and information.
With the current growing interest in Artificial Intelligence, there is the need to bridge the gap between Artificial Intelligence and Intellectual Property in Ghana. As Artificial Intelligence becomes widely used in Ghana, it is capable of generating creative works, inventions and even business solutions. Without clear Intellectual Property laws that defines who owns the rights, there can be legal uncertainties.
Ghana’s current Intellectual Property laws are governed by the Copyright Act, 2005, the Patents Act, 2003 and the Trademarks Act, 2004. These laws are aligned with international standards like the World Intellectual Property Organization (WIPO) treaties and the Agreement on Trade Related Aspects of Intellectual Property Rights.
These existing laws are designed to protect traditional forms of Intellectual Property, such as literary works, inventions and trademarks. However, they do not expressly address the unique challenges posed by Artificial Intelligence, such as ownership of AI generated works and the protection of algorithms.
In Ghana, there is currently no established regulatory framework for the general use of Artificial Intelligence. However, given the widespread integration of Artificial Intelligence across various industries, the necessity for developing rules and regulations to govern its use becomes imperative to prevent potential harm or adverse outcomes.
The UNESCO report during the launch of the UNESCO AI Ethics tool, highlighted that the Ghanaian government has expressed a keen interest in establishing a policy and regulatory framework for Artificial Intelligence within the country. This holds particular significance in Ghana, especially considering that in 2019, Google inaugurated its first AI office in Africa.
Even though there are no express AI terms used in the Copyright Act, 2005, it makes way for AI to thrive while protecting its users thereof.
Section 76 of the Copyright Act, 2005 (Act 690) defines literary work to include computer programs, whether in source or object code. This classification ensures that computer programs and software are treated as literary works under the law, granting them the same protections as other forms of written content such as books or manuscripts. By including computer programs in the definition of literary works, the Act ensures that the creators of these programs have exclusive rights to reproduce, distribute or modify their work, among other protections.
The recognition and protection of computer programs and software under the Copyright Act in Ghana provides a fundamental legal framework that can be extended to certain aspects of Artificial Intelligence. Here is how this existing protection could merge into protecting Artificial Intelligence in Ghana:
- AI systems as computer programs
Many Artificial Intelligence systems are fundamentally software programs, often complex ones involving algorithms, machine learning models and data processing techniques. Since the Copyright Act, 2005 already protects computer programs, some AI software could fall under this protection. This means the creators of AI software in Ghana have exclusive rights to their work, just like developers of other types of software.
- Protection of source and object code
Artificial Intelligence programs, whether in source code (the human readable code) or object code (the machine-readable code) are protected under the Copyright Act, 2005. This includes the underlying algorithms and code that drive AI applications, ensuring that these cannot be copied or modified without permission.
- Protection of AI generated outputs
A significant question in Artificial Intelligence and Intellectual Property law globally is whether AI-generated content such as art, music, text etc can be copyrighted. Under the current law in Ghana, copyright typically applies to works created by humans.
However, if an AI tool is used to assist a human in creating a work, can the human user hold copyright for the final output? How the courts in Ghana might interpret the protection of AI- generated works is still an open question. However, the existing legal framework can be adapted to extend protection to works that involve significant human input alongside AI assistance.
Section 1, Copyright Act, 2005 (Act 690)
“Work eligible for copyright
(1) An author, co-author or joint author of any of the following works is entitled to the copyright and protection afforded to that work under this Act; (g) derivative work”
Section 76 (interpretation Section) defines derivative work as;
“derivative work” means a work resulting from adaptation, translation or other transformation of an original work in so far as it constitutes an independent creation;
To qualify for copyright protection, a derivative work must demonstrate a level of originality and creativity that sets it apart from the source material, effectively making it an independent creation in its own right.
To what extent does AI generated content count as a derivative work?
One could certainly make the case that a work made by a human in concert with Generative AI, such as human artwork enhanced with AI, or AI text, edited and revised and published by a human could merit protection.
Or is the protection of a derivative work, flow only as far as the human being’s contribution to the work; the specific elements of the work that were injected by the hand of the human being?
Perhaps more central than any of the above is the argument made by anti-AI proponents, that the legislation never contemplated or anticipated a time when machines could autonomously generate artistic works, and therefore, copyright cannot subsist to a work that is not made by human actors.
Intellectual property challenges in the age of artificial intelligence
The need for legal clarity around AI-generated content is highlighted by the case of Zarya of the Dawn, a graphic novel created by Kristina Kashtanova, who used the AI tool, MidJourney to generate images for the work. Kashtanova sought copyright protection for her graphic novel, but the legal question quickly arose: Can AI-generated images, with minimal human intervention, be considered original works eligible for copyright?
The U.S. Copyright Office ultimately ruled that while the arrangement of text and the creative choice of how the images were placed within the story could be protected, the individual AI-generated images themselves were not eligible for copyright protection.
The decision hinged on the degree of human authorship involved. Kashtanova wrote the text and made choices regarding how to arrange and select the images, demonstrating that a significant level of human input was required to bring the story to life. However, because the images themselves were generated by MidJourney, a tool largely operating independently based on prompts, the Copyright Office deemed them ineligible for protection as they did not involve enough human creativity to qualify as original works.
To prevent legal ambiguity and ensure creators in Ghana are adequately protected, amendments to the Copyright Act, 2005 are essential. These amendments should clearly define when a human’s use of AI amounts to authorship and outline the level of creative input required for copyright protection, similar to the approach seen in the Zarya of the Dawn case.
The Patents Act, 2003 (Act 657) does not explicitly address whether inventions created or significantly enhanced by Artificial Intelligence are patentable, nor does it clarify how to assess the novelty, inventive step and industrial applicability of Artificial Intelligence driven innovations.
Guidelines are needed on how Artificial Intelligence contributions to inventions are treated under Patent law, including whether Artificial Intelligence can be listed as an inventor and the criteria for evaluating the inventive step when Artificial Intelligence is involved. There needs to be an introduction in the Patents Act, 2003, specific provisions that define the patentability of Artificial Intelligence generated inventions, addressing how Artificial Intelligence involvement impacts the evaluation process.
Artificial Intelligence systems rely heavily on large datasets, which often include proprietary or personal data. The current Intellectual Property laws in Ghana do not fully address ownership, usage rights and liability issues related to data used by Artificial Intelligence systems.
Legal Provisions should be established to clarify who owns the data used by Artificial Intelligence systems and the rights and obligations related to data sharing, especially when used to train AI models. There needs to be an update to the Intellectual Property laws to explicitly address data ownership, sharing and usage in the context of Artificial Intelligence, possibly introducing new regulations that govern AI data usage and rights.
Artificial Intelligence systems rely heavily on large datasets to train and refine their models, and a significant portion of this data is often sourced from the internet, including social media platforms. These platforms host vast amounts of user-generated content which are photos, videos, text, and metadata that can be valuable for AI training. However, the legal framework in Ghana currently does not address critical questions related to the ownership, usage rights, and liability for such data.
A discussion of data ownership in AI would be incomplete without considering the licensing agreements imposed by social media companies. Although Ghanaian laws might not explicitly regulate how data from these platforms is used for AI training, social media companies’ terms of service play a pivotal role in determining what is legally permissible. For instance, Facebook’s Terms of Service grants the platform a non-exclusive, transferable, sub-licensable, royalty-free license to use any content that users upload. This includes the right to sublicense the data for purposes like developing algorithms or training machine learning models.
Since Ghanaian law comprises common law (Article 11 of the 1992 Constitution) and by extension, the freedom to contract, one could make the case that these various licensing agreements with these internet companies that lease out our data to train these models have a sort of indirect, tenuous permission granted by Ghanaian law.
However, the counter case would be that such a game-changing technology is much too critical to be left to the whims of massive corporations and their nebulous terms of service agreements.
Ghana’s existing legal framework does not clearly address liability issues related to Artificial Intelligence generated content, particularly when Artificial Intelligence outputs result in Intellectual Property infringement, defamation, or other legal violations. It is essential to establish who is liable when AI-generated content infringes on Intellectual Property rights. This is a crucial gap, as AI systems are increasingly being used to generate content that can mirror or replicate existing creative works, leading to potential IP disputes.
A reasonable argument can be made that the liability should rest on the tech companies that create the AI software. These companies control the development and training of the AI systems, using vast datasets, often pulled from the internet without necessarily obtaining explicit consent or licensing for every piece of data. This is particularly pertinent when considering cases like New York Times v. OpenAI and the Microsoft OpenAI copyright infringement lawsuit.
In New York Times v. OpenAI, the New York Times filed a lawsuit against OpenAI, claiming that its models, such as ChatGPT, were trained using New York Times’ articles without appropriate licensing. The lawsuit emphasized the risk of AI systems unlawfully replicating and distributing copyrighted materials, and it sought to establish accountability for the use of proprietary content in AI model training. In that case, OpenAI was held accountable for how its technology utilizes copyrighted material, bringing to light the need for clearer laws regarding the relationship between tech companies, AI-generated content, and IP ownership.
Similarly, the Microsoft OpenAI lawsuit filed by authors against Microsoft and OpenAI centers on claims that the AI systems were trained using copyrighted works without permission, resulting in AI-generated outputs that violated IP laws. The lawsuit argues that because these companies designed and deployed the AI models, they are directly responsible for the infringement. As of now, there is no final decision in this case. There needs to be an introduction of legal provisions that outline liability for AI-generated content, balancing accountability between developers, operators, and potentially the AI system itself clearly.
Ethical considerations
As Artificial Intelligence continues to proliferate across industries, the intersection of Artificial Intelligence and Intellectual Property rights has become an increasingly complex and critical legal landscape. The rapid advancement of Artificial Intelligence has given rise to numerous ethical considerations and dilemmas in its relation to Intellectual Property.
The question of authorship and ownership of AI-generated works poses a profound ethical dilemma. Traditional Intellectual Property laws recognize human authors as the rightful owners of their creations. However, when an Artificial Intelligence system generates a piece of art, music or even software, the concept of authorship becomes ambiguous. The ethical challenge is determining who should hold the rights to such creations. Without clear legal guidelines, this could lead to disputes and ethical dilemmas regarding ownership and compensation.
Artificial Intelligence systems are trained on datasets that are copyrighted. There could be a future where individuals voluntarily opt into datasets with the knowledge that they will be used for AI training, or a community of people contribute to a publicly available dataset such as LAION-5B, which is the popular open source dataset.
These datasets may contain biases, leading to the creation of content that could perpetuate stereotypes or unfairly represent certain groups. This raises ethical questions about fairness and equality, especially when the content is protected under Intellectual Property laws. If AI-generated works are biased, they could contribute to the marginalization of certain communities or perpetuate cultural misrepresentations.
Artificial Intelligence algorithms can inadvertently generate content that infringes on existing copyrights. Ethical considerations arise concerning the responsibility of AI developers and users to address such infringements.
The Importance of Ethically Sourced Data in AI Development Artificial Intelligence algorithms, especially in creative domains, often generate content that may inadvertently infringe on existing copyrights. This raises crucial ethical questions regarding the responsibility of AI developers and users in ensuring that AI systems do not contribute to intellectual property violations
The rapid advancement of Artificial Intelligence has enabled the generation of hyper-realistic images, including deepfakes and explicit sexual content, which raises significant ethical concerns. The ability of AI systems to create, manipulate, and disseminate such content without consent poses unique threats to privacy, human dignity, and legal accountability.
Cases from around the world, including South Korea’s deepfake crisis, as well as incidents involving the use of AI to generate child pornography, highlight the urgent need for ethical and legal frameworks to govern AI-generated sexual images.
In South Korea, the prevalence of deepfake pornography, AI-generated sexual images that are manipulated to feature the faces of non-consenting individuals has sparked a national crisis. This phenomenon demonstrates the dark side of AI’s creative capabilities, where the technology is being used to invade privacy and objectify individuals, often women.
According to a report by NDTV, South Korea is one of the countries grappling with a surge in these AI-generated sexual images, as they disproportionately target women and girls, often for malicious purposes, such as harassment, blackmail, or humiliation. While the country has implemented laws to tackle this issue, the rapid pace of AI development continues to outstrip the capacity of these regulations, leaving victims vulnerable
In the realm where Artificial Intelligence intersects with Intellectual Property, a complex tapestry of ethical considerations unfolds. Through the lens of transparency, the mitigation of biases, the safeguarding of privacy and the cultivation of ethical consciousness, the ethical quandaries woven into the fabric of Artificial Intelligence and Intellectual Property can be navigated with prudence. By embracing these principles, the power of these technologies can be harnessed for the collective good.
Case studies
Case Study 1: The “DABUS” Patent Controversy
In 2021, an intriguing patent controversy emerged when Dr. Stephen Thaler submitted patent applications for two innovations created by AI systems known as “DABUS”. Dr. Thaler argued that DABUS should be recognized as the inventor, challenging conventional notions of inventorship. However, the U.S. Patent and Trademark Office (USPTO) rejected the applications, asserting that only natural persons can be inventors under current patent laws.
This case reflects the ongoing legal and ethical debate surrounding AI-generated inventions. The rejection by the USPTO highlights the gap between existing patent laws, which primarily consider human inventors, and the increasingly autonomous capabilities of AI. Addressing this issue is paramount to encouraging AI innovation while ensuring proper attribution and protection of intellectual property rights. The case underscores the need for a legal framework that accommodates AI’s role in invention and creatorship while preserving the principles of fairness and equity.
Case Study 2: The “NFT Art Generator” Dilemma
The rise of non-fungible tokens (NFTs) has reshaped the art world, and it was not long before AI became involved. In a notable instance, an AI algorithm was programmed to create digital art pieces, which were then sold as NFTs in online marketplaces. This development sparked debates concerning the ethical and legal aspects of selling AI- generated art, including questions about copyright ownership, attribution, and the rights of traditional artists who might have influenced the AI’s algorithms.
This case exemplifies the ethical complexities of AI- generated content within the realm of intellectual property. It accentuates the importance of adapting copyright and IP laws to address AI’s creative capabilities.
Finding a balance between AI innovation and protecting the rights of both AI developers and traditional artists is a challenging but essential task. This case underscores the need for transparent guidelines and legal mechanisms that ensure appropriate recognition and compensation for creators while fostering innovation in the evolving landscape of AI- generated art.
In both cases, there is a pressing need for legal frameworks that can adapt to the evolving role of AI in intellectual property. Current laws and regulations often fall short in addressing the unique challenges posed by AI-generated content and inventions.
Research in this area should explore potential legal reforms and policy changes to establish a more equitable and responsible approach to AI and IP. Additionally, these case studies underscore the importance of transparency, accountability, and ethical considerations in AI development and usage, as they can significantly impact IP rights and creative industries.
Recommendations and conclusion
The Copyright Act, 2005, along with broader intellectual property laws, must be updated to address these emerging complexities. A noteworthy precedent was set by the Writers Guild of America (WGA) in its 2023 agreement, which requires studios to disclose the use of AI in content creation.
This regulation acknowledges the increasing role of AI and aims to protect the rights of human creators, ensuring that AI is used as a tool rather than a replacement for creative talent. Such measures could serve as a model for other sectors as AI-generated content becomes more prevalent, stressing the need for transparency and accountability.
As Ghana stands at the cusp of a technological revolution driven by Artificial Intelligence, the intersection with Intellectual Property law presents both opportunities and challenges. The current landscape reveals substantial gaps in how Intellectual Property laws address AI-generated content and inventions.
Ghana’s existing statutes, including the Copyright Act, Patents Act, and Trademarks Act, while robust in the traditional context, fall short in accommodating the unique complexities posed by AI. The need for a comprehensive regulatory framework that addresses AI-specific issues such as the ownership of AI-generated works, the patentability of AI-driven inventions, and data rights has never been more urgent.
As Ghana continues to embrace AI, it is imperative that stakeholders—including policymakers, legal professionals, and technologists—collaborate to craft a flexible and forward-thinking legal framework.
This will not only protect the rights of creators and innovators but also enhance Ghana’s position as a leader in AI technology on the global stage. In bridging the gap between Artificial Intelligence and Intellectual Property, Ghana has the opportunity to create a legal environment that supports technological advancement while ensuring fairness and equity.
By addressing the legal and ethical challenges identified, Ghana can pave the way for a future where AI and IP harmoniously coexist, driving growth, creativity, and innovation.
>>>the writer is a Legal Intern at Koranteng & Koranteng Legal Advisors. Contact: [email protected]
