This Article is written by Pranjal Padmgirwar, a Fourth Year Law Student of Manikchand Pahade Law College, Chh. Sambhajinagar

Introduction

Artificial intelligence is disrupting everything, from healthcare to finance, from entertainment-there appears to be nothing that artificial intelligence does not "change" at light speed. The impact, however, of artificial intelligence on intellectual property law raises specific questions of challenges and opportunities that the global legal system must grapple with. As artificial intelligence technologies advance, existing frameworks surrounding intellectual property rights have never been under as much pressure as they are at present.[1] This essay digs into how AI will affect IP law and considers the possible openness of the current structures of the law toward such changes brought about by AI-generated works, difficult enforcement aspects, and the possibility of reform.

Understanding Artificial Intelligence and Intellectual Property

What is Artificial Intelligence?

Artificial intelligence covers many technologies: the ones that will be able to make machines perform what, up until then, was only conceived of for humans who had a good amount of intelligence. This would include learning from experience and patterns, understanding natural language, and making decisions. The three most applied approaches of AI dominating contemporary practice are machine learning, deep learning, and natural language processing.

Intellectual Property Law Basics

Intellectual property law is enacted so that intellectual productions regarding inventions, designs, brands, and artistic works would be protected by the exclusive rights over the creations of the creators.[2]. There are two categories of IP, namely:

1. Copyright: The copyright branch protects original works of authorship; be it literature, music, and all other visual arts. It grants the creator of that work an exclusive right to reproduce and distribute his work and show it publicly.

2. Patent: Refers to inventions and processes. It gives the property right of the invention or process to the inventor for a specified number of years, usually 20 years, to work on the invention and also commercially exploit it as well.

3. Trademark: Has reference to symbols, names, and slogans used to identify goods and services. This enables consumers to distinguish one brand from another.

4. Trade Secrets: This refers to confidential business information that is used for the betterment of a company vis-à-vis its competitors. It may be in the form of a formula, practice, design, etc.

Interaction of AI and IP Law

One of the most basic issues that AI poses for IP law is authorship. Typically, one of the underlying tenets of copyright law is that something cannot be protected if it does not have an author.[3]. However, can AI systems create music, art, literature, and other kinds of creative output independently, without any human intervention? This raises foundational questions regarding who owns rights to AI-generated works:

1. Authorship by the Creator This principle recognizes human creators of AI systems or rather those designing the algorithms and training the models to get them to work as authors of the works produced by their AI.[4] Here, the AI system is considered a tool as in photography about a camera or painting about a paintbrush, and ownership of the output goes to the creator.

2. The AI as a Separate Legal Entity: There are quite a few arguments that as the AI achieves greater levels of autonomy, it should be accorded some kind of legal personhood where it can become the owner of copyrights of its creations. This would be an extreme challenge to the bedrock of copyright law and raise further questions regarding liability and enforcement.[5]

3. User Contributions: Every time a user inputs data or triggers to order AI-generated content, one has to then consider the level of contribution a user has in his generated work. Should the developer of the AI own a portion of it with the user if he has dramatically altered the output?

Obstacles to Copyright Law

The emergence of AI-generated works throws a wrench into copyright law:

1. Originality and Fixation: Copyright protection is typically available only to those works that are original and fixed in a tangible medium. The question of determining originality, however, would thus be a concern as AI algorithms often generate their output from pre-existing data. Balance may now be confronted by courts on the issue of whether an AI-generated work can qualify for the originality standard.

2. Infringement Problems: The AI models appear to learn and generate new content based on gigantic datasets. Based on this approach, copyright infringement then appears as it has been making copies of works that are protected without the AI knowing the facts. However, there is no conclusion in the sense that the use of copyright materials in training AI models might be legit or not.[6]

3. Collective Works and Fair Use: A tough question arises with the use of AI to create collective works, such as a collection of music or an ensemble of artwork right for fair use. The courts will have to determine whether the use of copyrighted material by AI in creating such works can be transformed and thereby falls within the ambit of the fair use doctrine.

Patent Law and AI Inventions

Patentability of AI Innovations

This brings the discovery of new inventions and processes. Patent law is one tool used to encourage innovation and provide rights to make exclusive inventors. However, at the confluence of AI, issues are emerging:

1. Who is a human inventor? Traditionally, the law of patents requires the name of a human inventor to be disclosed in the patent application. This brings interesting questions in terms of who needs to be identified as an inventor: the creator of the AI system, the user, or the AI itself?

2. Novelty and Non-Obviousness: Patent law also requires that an invention be novel as well as non-obvious to a person having ordinary skills in the area. It is difficult to determine non-obviousness in AI-generated inventions, as the results generated by AI appear inventive but only due to known knowledge based upon known patterns.

3. Disclosure requirements: Patent applications should sufficiently disclose the invention to allow others to understand and reproduce it.[7] It becomes very difficult to fulfill disclosure requirements as AI algorithms are opaque and impossible to detail how an AI system has arrived at a particular invention.

SECTION-B: PROBLEMS OF ENFORCEMENT

In the context of AI, added challenges come into play when there is enforcement of IP rights. In AI technologies, identifying what examples of infringement look like becomes a challenge. For example, if one has an AI-generated work, much input might be required on the underlying algorithms and training data in order to classify whether the work infringes existing copyrights.

This is because AI developments are international, meaning that enforcement is hard to establish. IP laws are different in each country, which leads to the law being interpreted differently and, therefore, inadequate protection for the creators. This, therefore, poses a problem when organizations wish to protect innovations resulting from AI worldwide.[8]

The Case by the Legal Profession for Change

Because problems emerge with the dominant AI and the current IP framework, there is a feeling among most legal experts that it needs to be reformed in such a manner that these issues are better dealt with. Such possible reforms include:

1. Copyright Laws Reformed: For instance, the legislators might just have an idea of reforming the copyright statute to specifically advocate for AI-generated works by making clear authorship, ownership, and tests of originality. This would be in the form of joint authorship by the developers and users of an AI system and directives on the use of copyrighted material for training an AI system.

2. Introduction of new classes of patents: A new class of patents for inventions out of AI could be developed, which may thus pave the smooth way for patenting and even inventorship issues with correct inventorship based on the human aspect of AI inventions.

3. International Standard Development: AI technology is borderless, and the protection of rights should be harmonized under international standards, even though national legal standards could merge into one harmonized framework through international cooperation and mutual collaboration, responding to the issues AI raises in the IP regime.

4. Promotion of Ethical AI Development: Apart, from outside reform, it should be encouraged to develop a culture of ethical AI development. An encouraging aspect related to the use of algorithms and dataset transparency can help counter infringement concerns against AI technologies and increase public confidence in the same.[9]

Conclusion

Artificial intelligence deeply and multi-facetedly impacts intellectual property law. Without a doubt, embracing the evolved AI technologies will throw interesting challenges and opportunities that have to be addressed within legal systems. Protection of creators' rights while encouraging innovation would mandate willingness on the part of lawmakers and practitioners to adjust and redefine the existing frameworks of IP. If responded to promptly, it would certainly stand lawmakers and practitioners in good stead to shape the new frontier of AI and intellectual property as well as one could hope for society. The road toward a fair and efficient IP regime in the age of AI is still just beginning to be traveled, and more frequent and open communication between stakeholders is urgently needed for the future to cast this journey in a positive light.

References

[1] Christopher Heath and Anja M. Schartau, Intellectual Property Law in the Digital Age (Springer 2020).

[2] Sara E. E. L. Connolly, 'Copyright and Artificial Intelligence: Can AI Generate Copyrightable Works?' (2019) 40(1) European Intellectual Property Review 1.

[3] David I. Bainbridge, Intellectual Property (8th edn, Pearson 2019).

[4] R. Anthony Reese, 'Artificial Intelligence and the Future of Copyright' (2020) 94 Washington Law Review 163.

[5] World Intellectual Property Organization, 'Artificial Intelligence and Intellectual Property' (WIPO 2019) <https://www.wipo.int/edocs/mdocs/mdocs/en/wipo_ai_19/wipo_ai_19_6.pdf> accessed 30 September 2024.

[6] European Commission, 'A European Approach to Artificial Intelligence' (European Commission 2020) <https://ec.europa.eu/info/sites/default/files/commission-white-paper-ai-2020.pdf> accessed 03 October 2024.

[7] United States Patent and Trademark Office, 'Public Views on Artificial Intelligence and Intellectual Property Policy' (USPTO 2019) <https://www.uspto.gov/sites/default/files/documents/USPTO_AI_IP_Report.pdf> accessed 03 October 2024.

[8] Burrow-Giles Lithographic Co v Sarony 111 US 53 (1884).

[9] Yahoo! Inc v La Ligue Contre Le Racisme Et L’Antisemitisme [2004] ECR I-10981.