The Age of AI-Generated Content: Reshaping Intellectual Property Norms

The Hague’s Mauritshuis Museum has loaned "The Girl with the Pearl Earrings" to another museum and hung "A Girl With Glowing Earrings" (AI assisted creation) of Julian van Dieken as a substitute.

There is no doubt that artificial intelligence (AI) stands out as the hottest topic in the realm of technological development. Intellectual property (IP) systems have traditionally aimed to encourage human innovation and creativity and until very recently such qualities were deemed as human-specific characteristics. As artificial intelligence (AI) becomes increasingly integral across various sectors of the economy and society, it prompts fundamental questions central to existing intellectual property (IP) systems. At the forefront of these questions is whether products generated by artificial intelligence can be protected under intellectual property rights, and if so, who would be the owner of these IP rights. Additionally, it raises inquiries about whether these technological developments necessitate new legislative changes in intellectual property law.

This memo highlights the evolving challenges at the intersection of artificial intelligence and intellectual property rights, also emphasizing the need for legal adaptation to safeguard innovation in this rapidly changing landscape.

A Definition of Intangible Assets and Autonomous AI

Intellectual property rights constitute intangible assets that are subject to ownership, acquisition, and transfer but these assets lack physical substance. Intellectual Property Rights encompasses trademarks, designs, copyrights, trade secrets, patents, and various other rights. Intangible assets potentially generated by autonomous AI includes cutting-edge technology, software, artistic works, confidential information, and other elements, all of which would be safeguarded by intellectual property rights if created by a human.

How Do Generative AI Systems Do It?

Artificial Intelligence (AI) involves the creation of computer systems with the capacity to execute tasks that traditionally demand human intelligence. These systems acquire knowledge from data, identify patterns, make decisions, and adjust to new information. The learning capability of these systems empowers AI to autonomously execute intricate tasks, often requiring minimal human intervention. This differs Generative AI systems from Traditional AI because Generative AI Systems emphasizes creativity and the generation of diverse, often novel, content across various domains, while Traditional AI is designed for specific task-oriented functions with predefined solutions.

Nonetheless, this training procedure might raise concerns about intellectual property infringement, as it frequently involves the utilization of data scraped from the internet, potentially encompassing copyrighted content. On the other hand, questions surrounding intellectual property (IP) in AI go beyond protected data and training images to encompass the creative outputs. Addressing copyright, patent, trademark issues, and determining ownership of AI-generated content becomes crucial, particularly given the challenge that copyright requires human originality, which may be lacking in AI-generated works.

Can IP Rights Protect AI Systems?

a) Patent Protection

Despite intellectual property law evolving in response to artificial intelligence developments, recent legal statements emphasize that AI does not possess the legal status of a "person" within the frameworks of copyright and patent law. There is a global trend of courts and patent offices rejecting patent applications naming AI systems as inventors, emphasizing the consensus that an inventor must be a human or a person with legal capacity. And yet it raises the question of ‘‘If not AI, then who?’’ Governments globally are actively contemplating how the patent system should adjust to accommodate inventive AI contributions, proposing various reform options. Recognizing the tension between current patent laws and the dynamics of modern innovation, there is an emphasis on continuous adaptation to ensure that the patent system effectively encourages and rewards innovation.

b) Trademark Rights

In the realm of intellectual property, the essence of a brand lies in its ability to distinguish itself, prompting businesses to employ various representations to set their products apart. When utilizing AI for brand creation, businesses entrust the AI with translating their identity into tangible trademarks, reminiscent of human designers. However, the question of AI or its creator's rights over the resulting trademark has sparked legal debate. Traditionally, most legal systems don't recognize AI as a legal person, implying it can't claim ownership. Yet, a recent decision by the U.S. Copyright Office challenges this notion, allowing certain AI-generated artwork to be trademarked. This development hints at a shifting landscape, underscoring the need for businesses to navigate evolving legal frameworks as they engage AI in intellectual property creation.

c) Copyright

Copyright protection, in addition to patent protection, can extend to certain elements of AI systems, particularly their source code. Unlike patents, copyright arises automatically upon the creation of the protected material, eliminating the need for registration and reducing upfront costs. However, it's essential to recognize that while copyright safeguards the code, it doesn't protect the system's functionalities; instead, it focuses on the "form of expression" without safeguarding the underlying ideas. In the context of AI or machine learning systems, copyright infringement is not limited to verbatim copying of the source code. In the EU, while the structure and organization of a work may be eligible for copyright protection, the application of these principles varies among member states due to a lack of comprehensive harmonization in EU copyright law.

d) Confidential Information and Trade Secrets

Enhanced protection may be available through laws governing confidential information or trade secrets, extending to the operational aspects, training data, and algorithms of AI systems. In the EU, trade secrets protection is widely harmonized and ensures protection against unlawful misappropriation, provided adequate confidentiality measures are implemented. However, reliance on confidential information or trade secrets demands vigilant protection, as once information becomes public, restraining further unauthorized use becomes challenging, emphasizing the importance of carefully choosing practical protection routes for AI and ML systems.

In Lieu of Conclusion: Intellectual Property Law Needs to Catch Up

In April 2021, the European Commission introduced a proposal for an EU AI Act; adopting this act is still a long way off, and the AI Act Proposal needs to mention IP rights. We will explore the EU's stance on the intersection between AI and IP Rights.

The United States does not have a specific comprehensive federal law that solely addresses the intersection between artificial intelligence (AI) and intellectual property (IP) rights. On October 30, 2023, White House issued an Executive Order on Safe, Secure and Trustworthy Artificial Intelligence. The executive order emphasizes U.S. commitment to responsible AI innovation, directing investments in education, training, research, and development. It mandates the issuance of guidance on AI-developed inventions' patent eligibility and explores copyright protection for AI-generated works. The Secretary of Homeland Security is tasked with developing a program to mitigate AI-related IP risks.

In Turkish Law, as per the Law on Intellectual and Artistic Works of Art (FSEK), the creator is deemed the owner of a work, including processing and compilation, with due consideration to the rights of the original work owner. Article 1/B of FSEK defines a "author" as the person creating the work, and the 2004 amendment removed the term "real," extending rights to legal entities. This amendment has sparked discussions regarding the attribution of authorship to artificial intelligence in the context of products generated by AI.

Analyzing the registrability of artificial intelligence products under Turkish Law involves examining the eligibility for protection criteria outlined in the Industrial Property Law (SMK). Article 3 of the SMK specifies that protection is extended to citizens, real persons with registered addresses within the Republic of Turkey or legal persons engaged in activities therein, and citizens of the reciprocating countries. Notably, the legislation consistently refers to the author as a "person," posing a challenge to recognizing artificial intelligence as the author without legislative amendments. However, the claim of an individual presenting themselves as the author of an AI-generated product raises a distinct discussion.

Existing intellectual property laws do not satisfy the needs for issues related to potential intangible assets autonomously generated by artificial intelligence. Effectively addressing these issues necessitates collaboration among various stakeholders, with a particular emphasis on policymakers, service providers, developers, content creators, and business owners.

Please feel free to reach out for more information and further questions on IP Rights Protection of AI Technology. Don’t forget to follow us on LinkedIn to stay informed of our memorandums!