Artificial intelligence (AI) has evolved from sci-fi fantasy to modern technology in a blink of an eye. In today’s world, Apple Siri is ready to engage in a conversation whenever someone is feeling bored, the Google Assistant provides recommendations for the best stores to purchase one’s favourite dresses based on the browsing history and preferences, Facebook suggests travel destinations by analysing uploaded photos and locations and ChatGPT can literally assist you with any question that you have and is a Google on steroids.
These phenomena, driven entirely by AI rather than human intervention, highlight the undeniable presence and ubiquitous influence of AI in our daily lives. In the next decade, AI is expected to be a trillion United States Dollar industry, steering in what is commonly referred to as the ‘Industrial Revolution 4.0’ as highlighted by the World Economic Forum.
AI involves computers or robots performing tasks that require human-like intelligence, such as reasoning and learning. Since the 1940s, digital computers have excelled at complex tasks such as chess and mathematical proofs. While AI has not reached the full flexibility of human intelligence, it has achieved expert level performance in areas like medical diagnosis, search engines, voice recognition, and chatbots. Its’ most recent innovation, generative AI, has ushered in a completely new dimension in the field of AI altogether, where AI is now able to be used for a multitude of tasks which were not anticipated before.
As AI technology advances, its role in creating aesthetic works with minimal human intervention raises questions about authorship and ownership in the field of intellectual property (IP). AI systems, especially ones using generative AI, are now capable of generating works independently, prompting concerns over whether machines should receive the same rights as human creators. For example, Midjourney, based on the stable diffusion AI model, is a text driven image generation tool which can, with a simple textual prompt, generate the corresponding image in approximately one minute.
Just this year (in 2024), the winner of a Japanese literary prize, Rie Kudan explained that she used ChatGPT to write parts of her novel. X’s (previously Twitter) Grok chatbot now lets you create images from text prompts and publish them on X. These instances prove that AI can perhaps meet the standards of human creators in creating aesthetic works or even do better in certain instances, but, whether such works can attract copyright protection and the extent to which such protection can be exercised remains unsettled in law. This issue has thus gained importance due to the increasing use of AI-generated works, including in professional contexts, making the debate over IP rights more pressing and worth exploring.
Copyright protection of AI-generated works
The law of copyright has been in existence at least since the 16th Century in both civil and common law jurisdictions. Civil law in the French jurisdiction emphasises the author’s personality while common law, as illustrated in United Kingdom law, focuses on the protection of the work itself and rewards for labour. The core intention of copyright is to safeguard the intellectual efforts of the creator and prevent third parties from exploiting works without fair compensation to the original creator. Furthermore, the utilitarian approach supports granting protection to works that offer the greatest public benefit, incentivising innovation through limited monopolies that are balanced by constraints on scope and duration.
When traditional copyright theories are applied to AI-generated works, it could be suggested that AI-generated works should be excluded from copyright protection as copyright should reward human efforts and personal labour, and also because AI-generated works lack “personal expression” and a human touch. However, the incentive theory, highlighted in Sony Corp. of America vs. Universal City Studios Inc., argues that copyright aims to advance public benefit and creativity rather than merely protect human labour or personality. According to this theory, therefore, AI-generated works may be eligible for copyright protection as they promote innovation and the development of useful arts.
AI & the SL law of copyright
In Sri Lanka, IP law is governed by the IP Act, No. 36 of 2003, which replaced the Code of Intellectual Property Act, No. 52 of 1979. The Act aims to reorganise the registration, control, and administration of IP, including copyright. It protects the rights of authors over their literary, artistic, and scientific creations, ensuring that these original works are legally protected. Section 6 of the Act specifies that only original works which are created by an author, i.e., not copied from existing works, are entitled for protection. This approach to originality is based on principles of English law and European Union law, emphasising the importance of intellectual creation and the uniqueness of the expression.
The Act categorises copyright into economic and moral rights. Economic rights allow the owner to benefit financially from their work, while moral rights protect the author’s honour and reputation. Moral rights are of concern here as these rights enable authors to have their names or pseudonyms attributed to their works and to object to any modifications or uses that could harm their reputation. Therefore, these rights are meant and applicable specifically for human creators.
Copyright protection in Sri Lanka lasts for the lifetime of the author and another 70 years after the author’s death, which indicates that only human authors can hold such rights. Essentially therefore, since machines and AI lack legal personality in Sri Lanka, AI-generated works cannot be afforded the same protection. Consequently, Sri Lanka has not yet updated its laws to accommodate advances in AI technology. This is currently a discussion being had all around the world, and we will hopefully expect regulation in this arena pretty soon.
AI & neighbouring rights
In order for AI-generated works to become a reality, multiple parties are involved, all the way from software programmers to designers to end users. Therefore, certain academics have considered the possibility of conferring different rights based on the relative contributions that each party makes. Copyright law already makes provision for this concept under ‘neighbouring rights’ also known as ‘related rights’.
As per the Agreement on Trade Related Aspects of IP Rights (TRIPS Agreement), neighbouring rights are designed to protect those who contribute to the creation and dissemination of works, including through creative, technical, or organisational efforts. Currently, these rights are categorised into three groups to recognise the efforts of performing artists, producers of sound recordings, and broadcasting organisations. However, these rights have not yet been adopted for stakeholders of AI-generated works.
In the context of AI-generated works, the AI algorithm is initially programmed by a programmer, while the final work is created under the authority of the user. Post-programming, the data and information produced by the AI are generated by the users who operate the AI after purchasing it. As users create numerous works, the final output of the AI reflects its learning, which is influenced by the data and information provided by different users.
The purpose of neighbouring or related rights is to safeguard those who help bring works to the public and facilitate further productions. However, there is no statutory mechanism for the recognition of neighbouring rights for AI-generated works, in Sri Lanka or elsewhere. Consequently, until statutory reforms are implemented, it can be argued that Sri Lankan courts could potentially extend protection under neighbouring rights to relevant stakeholders including programmers and users of the AI, if they are deemed to have contributed to the creation of AI-generated works.
Reforming the SL legal framework
Due to the current lack of regulation in AI in Sri Lanka and across the world, it is crucial to draft and implement laws to address the challenges posed to IP by rapidly advancing AI technologies, which may soon exhibit human-like reasoning abilities. The inspection of various legal frameworks reveals that the granting of legal personality to AI or machines is impractical, as it differs significantly from the legal status of a company. However, recognising joint authorship rights may be a more suitable approach, where authorship rights could be granted to both programmers and users proportionally.
In Sri Lanka, the Government is in the process of addressing this shift by forming a National AI Strategy for 2024-2028, investing Rs. 1.5 billion, and establishing a National AI Centre. Despite improvements in AI readiness, current challenges of Sri Lanka include data infrastructure, human capital, and technology maturity.
In conclusion, Sri Lanka is in need of modernising its copyright laws to address the complications of AI-generated works and authorship rights. As AI technologies evolve faster, the existing legal framework must adapt to recognise the contributions of both programmers and users as well as the AI in creating AI-generated content. The implementation of a model of joint authorship or similar frameworks could provide a balanced approach, ensuring that all parties involved receive appropriate recognition and protection. By proactively updating its IP laws, Sri Lanka can better align with the proposed global standards and foster innovation in the rapidly advancing field of AI.
(The writer is an attorney-at-law and a legal practitioner in civil litigation)
……………..…………………………….
The views and opinions expressed in this article are those of the author, and do not necessarily reflect those of this publication