Who Gets the Credit? AI, Humans, & the Quirky World of IP Rights
Artificial intelligence (AI) has become an undeniable force in our daily lives, moving from a futuristic concept to a present-day reality. First conceptualized by Alan Turing as early as 1950, the term "artificial intelligence" itself was popularized in 1956 by John McCarthy and Marvin Lee Minsky, who established AI as a distinct field of research. At its core, AI refers to systems that act intelligently by observing their surroundings and taking actions, with a degree of independence, to achieve specific goals. This "intelligence" mimics the human or animal brain's ability to learn and solve problems, whether by imitating human thought (a cognitive approach) or through computational logic (a rationality approach). After periods of fluctuating interest, AI is now deeply integrated into our lives, powering everything from virtual assistants to semi-autonomous vehicles. Its incredible adaptability means it's being used to solve complex problems across various fields, including medicine, policing, and justice, by excelling at tasks like natural language processing, knowledge representation, automated reasoning, machine learning, computer vision, and robotics. Given this pervasive influence, it’s hardly surprising that AI profoundly interacts with the world of intellectual property (IP). This relationship is a two-way street, where IP influences AI and AI influences IP, and it can be both beneficial and challenging.
The intersection between AI and IP can be understood through three main areas: first, AI serving as a tool to help manage IP rights; second, IP laws being used to protect AI itself; and third, IP laws potentially creating barriers to the transparency of AI systems. Each of these areas presents unique opportunities and complex questions that are still being grappled with today.
Firstly, AI is proving to be an incredibly helpful technology in managing intellectual property rights. IP administration, which involves handling applications for protection, is increasingly deploying AI-powered solutions. A prime example of this is the World Intellectual Property Organization (WIPO), which uses AI-based applications like "WIPO Translate" for automated language translation and "WIPO Brand Image Search" for recognizing brand images. Beyond WIPO, many other IP offices around the world have developed and implemented their own AI applications to streamline their processes. In fact, WIPO organized a meeting in May 2018 specifically to discuss these AI applications, encouraging different offices to share their innovations and exchange information. This shows how AI is making the administrative side of IP more efficient and accessible, benefiting applicants and administrators alike.
Secondly, intellectual property serves as a legal system designed to protect AI itself. IP policy fundamentally aims to spark new ideas and creativity in economic and technological fields, which is precisely where AI and IP naturally connect. AI is already having a significant effect on how we create, produce, and distribute various economic and cultural goods and services. When it comes to protecting AI, patent and copyright laws are the most relevant systems. However, this area introduces some of the most complex and debated questions, especially concerning inventions and creative works that AI systems generate largely on their own.
A significant challenge arises when AI applications autonomously create inventions. This leads to fundamental questions about who should be recognized as the inventor.
Should an AI application itself be named as the inventor?
If not, then must a human being always be named as the inventor?
If a human is required, how should the law guide us in determining who that human inventor is? Or should this be left up to the parties involved to decide through private agreements?
Furthermore, a crucial question is who should legally own a patent for an invention that an AI application was involved in creating?
Beyond inventorship and ownership, there's also the question of whether current legal rules are enough to handle the unique aspects of AI-generated inventions, or if new, specific laws are needed. Some even raise the possibility that inventions created autonomously by AI applications should be entirely excluded from patent protection.
Another important aspect of patent law is "patentability," which includes concepts like the "inventive step" or "non-obviousness" – meaning an invention must not be obvious to someone skilled in the field. When an AI generates an invention, it becomes difficult to determine what existing knowledge or "prior art" this standard should be compared against.
Finally, the requirement for "disclosure" in patent applications becomes particularly challenging with AI-generated inventions. A patent application must describe the invention clearly enough for others to understand and potentially replicate it.
How can this disclosure requirement be met when the algorithms used in machine learning change and evolve over time?
Should the data used to train an AI algorithm be disclosed or described in the patent application? The answers to these questions could lead lawmakers to consider creating an entirely new, specialized legal system (known as a "sui generis" system) of IP rights specifically for AI-generated inventions, aiming to ensure that there are still incentives for innovation in AI.
Similar difficult questions emerge when considering how copyright law applies to AI. AI applications are now capable of independently producing original literary works and artistic creations. However, the current copyright system is deeply connected to the idea of human creativity and the need to respect and reward the expression of human imagination. This creates a central conflict:
Should copyright be granted to original literary and artistic works that are autonomously generated by AI?
Or must a human creator always be required for a work to receive copyright protection?
Thirdly, while there's a push to protect AI systems with IP rights, this also introduces a significant issue: IP rights themselves could become an obstacle to the transparency of AI systems. There is a growing demand for greater openness and accountability in systems that make decisions using algorithms. However, fulfilling this demand is tough when machine learning processes involve numerous data sources, are constantly evolving, and have parts that are naturally unclear due to technology or legal reasons.
Here, IP rights, especially trade secrets, can create barriers. This is where the overlap between AI and IP becomes most challenging and potentially conflictual. For instance, a company might claim that revealing how its AI system works would expose valuable trade secrets. However, the source clarifies that the goal of transparency and accountability often doesn't require disclosing the secret "algorithmic rules themselves," but rather explaining their results or how they arrived at a decision. In other words, simply explaining an AI's output typically doesn't risk breaching IP rights or trade secrets. Moreover, IP rights should not be used as an excuse to avoid explaining or disclosing information about AI systems. This transparency issue clearly illustrates a situation where the relationship between AI and IP is most fraught with conflict.
In conclusion, the intersection of AI and intellectual property is a dynamic and evolving landscape. We've seen how AI can be a valuable tool for managing IP rights, making processes like translation and image recognition more efficient. However, the use of IP laws to protect AI, particularly in terms of patents and copyrights for AI-generated works, raises profound questions about inventorship, ownership, and the very nature of creativity and innovation. Furthermore, the tension between IP protection and the growing demand for transparency in AI systems highlights a significant area of conflict, especially concerning trade secrets. These are crucial and largely unanswered questions that call for careful consideration by scholars and policymakers. Finding the right balance between fostering AI innovation and ensuring fairness, accountability, and transparency is a critical task for the future.
IP Researchers:
Dr. Agus Sardjono: Professor of Law at Universitas Indonesia, specializing in Indonesian Business Law and Intellectual Property Law. He has a PhD in Law from Universitas Indonesia and has pursued advanced IP studies at the WIPO Academy and the University of Turin. He also served as a research fellow at Tokyo University.
Dr. Min Chueh Chang: A reproductive biologist who co-invented the oral contraceptive pill, revolutionizing birth control. Born in China, Dr. Chang made significant contributions to the understanding of mammalian fertilization and in vitro fertilization (IVF), paving the way for millions of births through IVF procedures.
Dr. David Ho: A globally renowned virologist and physician, who made breakthroughs in HIV research, particularly with the introduction of combination drug therapy, significantly improving the lives of people living with HIV. His impactful work has been recognized globally, and he's also known for his role in co-creating YouTube.
