Overcoming Obstacles in AI-Related Patent Applications

AI poses challenging but surmountable questions regarding inventorship, written description, and enablement standards in patent law.

 n recent years, especially since the introduction of ChatGPT by OpenAI, artificial intelligence (AI) has demonstrated significant practical results, leading to the development of new products and services adopted across various industries at a rapid pace. AI technology has revolutionized automation by enabling the handling of complex and costly processes that were previously beyond the capabilities of computers. Customer service, supply chain management, and numerous other tasks traditionally done manually are now being automated with the help of AI. Additionally, AI-powered products and services have become more accurate and efficient, making them highly attractive to businesses seeking improved performance.

While much attention has been focused on large language models (LLMs) and their ability to provide natural language responses to queries, AI also facilitates innovation in diverse fields like pharmaceuticals, robotics, and autonomous vehicles. The use of AI to turbocharge innovation carries significant potential implications for the patent system.

AI contributes to the invention process by generating, evaluating, and filtering candidate inventions, such as candidate drugs, in Moderna's example. It is essential to note that AI does not replace human inventors; rather, it collaborates with them, forming a human-computer team that can generate superior inventions more efficiently than humans alone. AI builds upon the foundation of previous invention-facilitation tools like computer-aided design (CAD), 3D printing, and even conceptual tools like mathematical methods throughout history.

Integrating AI into the patent system poses challenges when understanding and describing AI systems and their outputs. This presents hurdles for those seeking to meet patent law's requirements for written description and enablement when drafting patent applications for AI-related inventions. While challenging, these obstacles can be overcome. In certain cases, chemical and biological arts inventors have managed to patent materials based on their function, even without fully understanding their underlying mechanisms.

Patent law provides various ways to address such situations, including product-by-process claims and method-of-use claims, which allow the patenting of products based on the processes used to produce them or their intended use, respectively, even if the underlying mechanisms are not fully known. Applying these practices to AI-related patents will require patent attorneys and examiners to draw from their experiences in the chemical and biological arts, even when dealing with inventions traditionally categorized within the computer-related arts.

While large language models have garnered considerable attention, the broader recognition of AI's vast applications in automating the inventive process is still evolving. Many in the general public and the legal profession are yet to fully grasp the extent of AI's impact on innovation. It is crucial to address these challenges proactively and deliberately rather than reactively to ensure the patent system continues to foster innovation. By doing so, industrialists can promote innovation based on a clear understanding of how today's inventions are created, harnessing the power of AI to drive advancements across various industries.