Artificial Intelligence and Intellectual Property in the Cayman Islands and the BVI

The rapid advancement of artificial intelligence (AI) continues to raise complex questions about the applicability of intellectual property (IP) laws to AI and AI-generated works.

IP remains one of the leading and most contentious issues in respect of AI governance. AI adoption continues to grow, and this year is already showcasing a wider range of commercial applications across all sectors. In light of the IP related challenges, businesses leveraging AI technologies must strategize to navigate the evolving intellectual property landscape. As the legal framework around AI continues to develop, businesses need to ensure that they avoid copyright infringement while also effectively safeguarding their own IP assets. Developing a clear understanding of how existing laws apply to AI technology and staying updated on legal developments will be crucial for companies seeking to innovate responsibly and protect their intellectual property in an AI-driven economy.

Copyright protection in the BVI (for “qualified persons”) and in the Cayman Islands (for “qualifying persons”) is automatic when such person creates an original work (once the work is recorded, in writing or otherwise) such as sound and music recordings; films; when you write a book or poem; or when you develop new software. By virtue of the Copyright (Cayman Islands) Order 2015 and Order 2016 (as amended), Part 1 of the U.K.’s Copyright, Designs and Patents Act 1988, subject to certain exclusions and modifications, was extended to the Cayman Islands. The BVI has implemented its own legislation in the form of the Copyright Act (Revised 2020) which is very similar.

Patent protection in the two jurisdictions is quite similar. As it is in other jurisdictions, in order to get patent protection, the invention must be: (i) new – i.e. the first in the world, (ii) useful – i.e. the invention must serve a purpose or provide a solution, and (iii) inventive – i.e. the invention must not be obvious to persons in the industry in which the invention is intended to be used.

Both the BVI and the Cayman Islands allow for the indirect registration of patents. Once a U.K. patent is granted, an application can be made in either of the Cayman Islands or the BVI to extend the scope of protection. In the Cayman Islands, there is no deadline for the filing of the application to extend rights, whereas in the BVI, rights must be extended within three years from the date of issue of the UK patent.

IP laws are designed to protect human creations. Generative AI (AI which generates text, images, speech, video or technical inventions based on user-inputted instructions) continues to increase in capability and grow in adoption. However, most copyright and patent laws, for example, do not yet explicitly address AI’s role in authorship or inventorship, leaving a legal void requiring attention.  Traditionally, the author or inventor is the person or organisation that creates the works. If now AI is responsible for content creation autonomously without any human input, the question is who owns the copyright protecting such content.

• For countries, such as the UK, this may be answered by the fact that computer-generated works will be owned by the person who made the necessary arrangements for the creation of the work.
• An overwhelming view in the E.U. is that AI cannot be a legitimate author. However, specific ways of using AI may result in a work that is protected for the user.

We expect to see governments across the world grappling with balancing strategies aimed at encouraging the development of AI and innovation while, at the same time, attempting to modernize IP and AI legal frameworks to account for AI.

Training generative AI involves using large bodies of IP-protected works/ data in ways that may be infringing under current laws. Governments seeking to “unlock” the potential of generative AI are now more often looking to legislate to permit text and data mining (TDM) of IP-protected data in order to train AI. The intellectual property in the data used to train AI models is growing as a subject of legislative discourse and is now a key issue in matters that have flooded courts across the world, whether use of copyright-protected materials to train AI models infringes copyright.

Training AI using personal data or protected IP also provides challenges to legislators worldwide. Over the next 2-3 years we expect to see increased regulatory scrutiny of companies that create or use AI technologies which have been trained using (i) personal data and/or (ii) information/data protected by IP rights. Regulators worldwide are now paying greater attention to balancing the benefits of AI against concerns about personal data and the protection of IP, and we expect that this will continue in the next few years.

AI programs usually qualify as IP with software or computer programs being literary works. In some countries however, copyright protection will not apply for functional aspects of AI such as algorithms or system designs. AI systems function however by processing human-provided instructions to generate problem-solving outcomes. This capability makes AI-based programs highly valuable from an IP perspective, as their innovative nature and diverse utility underline their significance of IP protection.

Who owns the copyright?

Copyright laws often require that there must be a natural person to whom copyright can be attributed and many jurisdictions including the Cayman Islands and the BVI, do not provide for “computer generated” works where no human author is involved. This creates a gap in the protection of AI-generated works, which are typically produced autonomously with little or no human intervention. Many copyright laws also require that “sufficient effort” must be expended to make any literary, musical, or artistic work original in character – which involves time, human labour, and skill. What constitutes sufficient effort for AI-generated content remains largely untested, raising debates about whether crafting prompts or editing AI output meets the thresholds. Additionally, if non-human entities are recognized as “authors” then copyright duration may become complex. Generally, copyright protection is granted for an author’s lifetime plus a period of time following, potentially leading to indefinite protection for AI-generated works.

The duration of copyright protection in the Cayman Islands varies depending on the nature of the work at issue. For example:

1. For Literary, Dramatic, Musical or Artistic Works: copyright expires at the end of 70 years from the end of the calendar year in which the author dies. However, if the author is unknown, copyright expires at the end of 70 years from the end of the calendar year in which the work was made or first made public
2. For Computer Generated Literary, Dramatic, Musical or Artistic Works: copyright expires at the end of 50 years from the end of the calendar year in which the work was made.
3. For Sound Recordings: copyright expires at the end of 50 years from the end of the calendar year in which the sound recording was made or first made public.
4. For Films: copyright expires at the end of 70 years from the end of the calendar year in which the film/movie was made or first made public.
5. For Broadcasts: copyright expires at the end of 50 years from the end of the calendar year in which the broadcast was made.

Copyright Protection and Deepfakes

Most IP laws are ill-equipped to address the challenges posed by digital replicas or deepfake technology. Copyright law generally is not fit for purpose in respect of deepfakes as the source material for many deepfakes either falls outside the scope of copyright protection or the copyright owner is not the individual who is harmed by the infringement. Possible causes of action presented by deepfakes include (1) copyright infringement (if a deepfake involves unauthorized use of copyrighted material), (2) trademark infringement (if it uses a registered trademark without permission), (3) the tort of passing off (if it misrepresents a product or service as endorsed by a well-known individual), (4) personal data privacy violations or (5) defamation (if the content defames an individual).  The issue of whether the outputs of AI models – particularly where they substantially reproduce source materials – may infringe copyright and who may be responsible for such infringement, is also unresolved.

Is the person who infringed: the user of the AI generated work without the rights holder’s consent or the AI developer or AI system owner? This ambiguity poses risks for businesses. If there is no clear proprietary right in AI-generated works, businesses may be exposed to unnecessary risk.

AI programs usually qualify as IP with software or computer programs being literary works. In some countries however, copyright protection will not apply for functional aspects of AI such as algorithms or system designs. AI systems function however by processing human-provided instructions to generate problem-solving outcomes. This capability makes AI-based programs highly valuable from an IP perspective, as their innovative nature and diverse utility underline their significance of IP protection.

Fair dealing exceptions

With respect copyright infringement, some jurisdictions such as the Cayman Islands, BVI, U.S., U.K., Australia, Hong Kong and Singapore provide fair dealing exceptions for particular activities, despite the fact that many other jurisdictions do not.

Under Cayman Islands law and BVI law, some of the fair dealing exceptions permitted are:

1. Personal copying for private use
2. Non-commercial research and private study
3. Text and data mining for non-commercial research
4. Criticising, reviewing and reporting current events
5. Use for parody, caricature and pastiche
6. Making backup copies, de-compilation, observing, testing and studying, and correcting computer programme errors.

Who is the inventor? Will the requirement for “novelty” remain?

Many patent laws require the inventor to be a natural person. This requirement could exclude AI from being independently recognized as an inventor. AI-driven innovations, such as those involving algorithms and machine learning processes, face challenges in meeting the criteria for protection as an invention. Under US patent law, for example, absent at least one human inventor, an invention is not patentable. Also, the criteria for patentability in many jurisdictions (including the Cayman Islands and the BVI) usually involves requirements for novelty, an inventive step and industrial applicability. These criteria raise questions about whether AI-generated inventions can ever meet the inventive step requirement, traditionally linked to human ingenuity. AI relies on algorithms and datasets to mimic human cognitive functions, enabling it to generate patentable inventions.

We expect to see a continuing updating of guidance as to the level and type of human contributions that are necessary to support patentability as new cases make their way through courts in various jurisdictions. A vital question that arises is whether AI can be regarded as a legitimate author of content that it generates or an inventor in case of patents, given the want of legal personality of the AI itself.

A particularly effective legal structure for an AI-driven entity is the Cayman Islands foundation company. Such foundation companies do not require shareholders, allowing them to function with a governance model that can be tailored to an AI’s decision-making models. Key advantages of using such foundation companies include the following.

• Legal recognition: foundation companies can provide a defined legal entity that can interact with traditional financial institutions, sign contracts and meet compliance obligations.
• Decentralised governance: the ability to structure the foundation company without shareholders allows for governance mechanisms that can adopt smart contract-based decision-making or AI-driven decision-making.
• Asset protection and Tax mitigation: foundation company can be tax resident in the Cayman Islands, hold and manage assets and ensure legal clarity in asset ownership.
• Regulatory compliance: foundation companies can be designed to comply with regulations, including AML and CFT requirements, making them suitable for global transactions.
• Economic substance rules: foundation companies limited by guarantee are specifically exempted from the economic substance rules and can therefore hold and even make profits from intellectual property without coming under the relevant economic substance compliance regime.

VIEW FULL PDF

This publication is not intended to be a substitute for specific legal advice or a legal opinion. For specific advice on the matters covered above, please contact your usual Loeb Smith attorney or any of the following: 

E: gary.smith@loebsmith.com
E: robert.farrell@loebsmith.com
E: elizabeth.kenny@loebsmith.com
E: kate.sun@loebsmith.com
E: vanisha.harjani@loebsmith.com
E: edmond.fung@loebsmith.com
E: vivian.huang@loebsmith.com
E: faye.huang@loebsmith.com
E: yun.sheng@loebsmith.com