Generative Artificial Intelligence (AI) is transforming the world as we know it. From healthcare to entertainment, from education to finance, AI is reshaping various sectors and industries with its unprecedented capabilities. AI tools can now generate a vast range of content, including text, video, images, and sound. AI can also assist or could even replace human inventors in creating novel and useful inventions.
These newly released features (new by way of human sentience) raise complex and challenging legal questions around intellectual property (IP) rights of AI-generated works and AI inventorship of patents. Who owns the rights to the content or inventions created by AI? How can we protect the interests of human creators and innovators who use AI as a tool? How can we balance the incentives for innovation and the public interest in the age of AI?
When AI becomes the actual content machine
AI-generated content refers to any content that is created wholly or partly by AI, without direct human input or intervention. Examples of AI-generated content include text, images, music, videos, and even cinematic language.
AI-generated content poses a dilemma for IP law, which is traditionally based on the notion of human authorship or creativity. IP law grants exclusive rights to the creators or authors of original works of expression, such as books, songs, paintings, etc. These rights include the right to reproduce, distribute, perform, display, and make derivative works of the original work. IP law also protects the moral rights of the authors, such as the right to be identified as the author and the right to object to any distortion or modification of the work.
However, AI-generated content challenges the assumptions and criteria of IP law, such as originality, creativity, and authorship.
Is AI-generated content original, if it is derived from existing data or works? Is AI-generated content creative, if it is produced by algorithms or machines? Is AI-generated content authored, if there is no human involvement or intention? Various jurisdictions are tackling these questions independently, potentially leading to a legal quandary regarding AI-generated content.
In the United States, the law has been interpreted to require human authorship for IP protection. In 2023, the US District Court for the District of Columbia reaffirmed in Thaler v. Perlmutter the absence of ownership for works generated by an artificial intelligence tool. This means that AI-generated content, in its current state, is considered to be in the public domain, without IP protection. However, if a human author has made creative arrangements or substantial modifications to AI-generated content, these could potentially be protected under IP law.
This position was made famous by the U.S. courts in the “monkey selfie” case, where a crested macaque named Naruto took a selfie using a camera that was “left unattended” by a photographer. The photographer claimed IP rights over the image while the argument for Naruto having copyright was also made, but the court ruled that a monkey cannot own IP rights under U.S. law. This case highlights the complexities to be faced when applying traditional IP laws to the rapidly evolving field of AI.
The U.S. approach has some advantages, such as preserving the human element in the creative process and avoiding the potential pitfalls of granting IP rights to non-human entities, such as liability, enforcement, and moral issues. However, it also has some drawbacks, such as leaving AI-generated content unprotected and vulnerable to exploitation, potentially discouraging the development and use of AI tools for content creation, and creating uncertainty and inconsistency for human creators and users of AI-generated content.
In Europe, the recent updates in copyright law have brought some clarity. The AI Act, which is expected to be adopted in the second half of 2024, has expanded its scope to include rules for general-purpose AI systems and models.
This is in response to the emergence of generative AI models such as GPT4 and Midjourney, Sora. The Act includes provisions that require providers of these models to respect copyright law and to make publicly available a detailed summary of the content used for training the model. This implies that the ownership of copyright for AI-generated content still resides with the human entities that provide and manage these AI systems. However, there is currently no specific operator liability at the EU level for copyright infringements in the area of generative AI, suggesting that the question of ownership in cases of infringement is still open to interpretation. The recent agreement on the world’s first AI Act by the European Union is a significant step towards addressing these complex issues.
When AI starts to act more like Albert AI-nstein
AI inventorship of patents refers to the possibility of granting patent rights to inventions that are created wholly or partly by AI, without human input or intervention. Examples of AI inventorship of patents include inventions that are generated, designed, or optimized by AI systems, such as new chemical compounds, medical devices, or software algorithms.
AI inventorship of patents presents a fascinating conundrum at the crossroads of technology and intellectual property law. IP law grants exclusive rights to the inventors and/or applicants of novel and useful inventions, such as products, processes, or methods. These rights prevent others from making, using, selling, offering for sale, or importing the patented invention. IP law also provides protections such as the right to be named as the inventor.
However, AI inventorship of patents challenges fundamentals of IP law, such as novelty and inventiveness. Can an AI system be credited as the sole inventor of a patent? How do we define novelty and inventiveness when AI algorithms draw from existing data or works? Both the US and European approaches aim to balance the recognition of AI’s role in innovation with the preservation of human-centric principles in IP law.
In the United States, the USPTO guidance on inventorship for AI-assisted inventions clarifies that while AI systems cannot be credited as inventors, inventions created with the help of AI are not automatically unpatentable. The focus is on the human contribution to the inventive process, ensuring that the individual who made a significant contribution to the conception of the invention is named as the inventor.
The DABUS case in the UK and Europe highlighted the challenges of AI inventorship. The UK courts and European Patent Office (EPO) held that an inventor must be a natural person, thus rejecting patent applications that named an AI system as the sole inventor. This stance underscores the importance of human involvement in the inventive process and the legal and ethical implications of attributing inventorship to non-human entities.
When AI algorithms create inventions by generating from existing data or works, the challenge lies in determining whether the AI has merely used what is already known or has indeed created something new and non-obvious. The distinction between AI-assisted and AI-generated inventions becomes crucial here. For AI-assisted inventions, where a human provides significant input, the novelty and inventiveness can be attributed to the human’s contribution. However, in a future with autonomously AI-generated inventions, the question becomes more complex. If the AI system can generate an invention that meets the requirements of novelty and inventive step, it could arguably satisfy the criteria for patentability.
The answers to these questions could provoke further issues such as the potential for AI’s ability to accelerate the rate of invention that could overwhelm the patent system, leading to a disparity between those who can afford AI systems and those who cannot. This could also alter the nature of invention itself, shifting from human ingenuity to machine efficiency.
When AI takes the stand
As AI becomes more advanced and ubiquitous, it will continue to pose new opportunities and challenges for IP law and policy.
One future prospect for AI and IP is the development of specialized AI-related IP laws or revision of existing laws, that can better accommodate the specific features and needs of AI as a content creator and innovator. Such laws could provide clear and consistent definitions and criteria for AI-generated content and AI inventorship of patents, as well as the allocation and exercise of IP rights. Such laws could also address the ethical and social implications of granting IP rights to AI, such as the responsibility, accountability, and morality of AI and its owners or users. We are seeing the beginning for regulation with the EU AI Act and guidance for governmental bodies such as the USPTO.
Another future prospect for AI and IP is the establishment of international standards for AI technology protection, that can foster global cooperation and collaboration among different jurisdictions and stakeholders. Such standards could provide common and consistent rules and guidelines for AI-generated content and AI inventorship of patents, as well as the allocation and exercise of IP rights. Such standards could also promote the development and dissemination of AI technology and its applications, as well as the ethical and social values and principles of AI. Some organizations, such as the World Intellectual Property Organization (WIPO) and the International Organization for Standardization (ISO), have already initiated the process of establishing international standards for AI technology protection.
However, establishing laws or international standards for AI technology protection is not without challenges. Such as requiring extensive research, consultation, and coordination among various stakeholders, such as policymakers, legal experts, technology developers, and users. They would also need to balance the interests and incentives of different parties, such as human creators, AI creators, IP owners, IP users, and the public. Such standards would also need to be adaptable and flexible, as AI technology and its applications continue to evolve and change.
AI and IP are two fields that are constantly evolving and influencing each other. AI generated content and inventorship of patents is a fascinating and challenging phenomenon that poses significant legal questions. Developing specialized AI-related IP laws and establishing international standards for AI technology protection are some of the future prospects. However, there are no clear or easy answers to the question of who owns what in the age of AI, yet. Collaborative efforts between policymakers, legal experts, and technology stakeholders will shape the future landscape of AI and IP. Nonetheless as we continue to explore the capabilities and implications of AI, we should also continue to value and emphasize the irreplaceable role of human inventiveness and innovation in the IP system.