Why AI Cannot Be Listed as a Legal Inventor on Patents

Patent offices and courts worldwide have consistently ruled that artificial intelligence systems, most notably the AI known as DABUS, cannot be listed as the legal inventor on a utility patent application. This global consensus is rooted in the strict statutory language of existing patent laws, which explicitly define an inventor as a "natural person" or human being. Beyond the literal text of the law, authorities reject AI inventorship because software lacks legal personality, cannot hold or transfer property rights, and cannot fulfill the human conception requirement essential to patentability.

The primary reason patent offices reject AI inventors is the plain language of patent statutes. In jurisdictions such as the United States, the United Kingdom, and the European Union, patent frameworks were written long before autonomous algorithms existed. For example, the U.S. Patent Act uses pronouns like "himself" and "herself" and explicitly defines an inventor as an "individual." Both the U.S. Patent and Trademark Office (USPTO) and federal courts have established that under standard legal interpretation, an "individual" refers strictly to a natural human being, leaving no room for machines to be recognized as inventors.

Another major hurdle is the legal concept of "conception." In patent law, conception is defined as the formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention. Because inventorship requires a mental act, patent examiners and judges argue that an AI system cannot legally conceive of an invention. Even advanced neural networks like DABUS do not possess a conscious mind; they process data and generate outputs based on human-designed parameters and training sets. Consequently, the law views AI as a sophisticated tool used by humans, much like a microscope or a laboratory computer, rather than an independent creator.

Furthermore, granting inventorship to an AI creates impossible legal complications regarding ownership and accountability. Patent applications require the listed inventor to sign an oath or declaration stating that they believe themselves to be the original inventor. Making a false statement on this oath carries legal penalties, including charges of perjury. Because an AI lacks legal personhood and moral agency, it cannot understand an oath, sign a legal document, or be held legally accountable in a court of law.

Ownership rights also flow directly from the inventor. An inventor initially owns the rights to a patent and can legally assign or license those rights to a corporation or another person. Because an AI is considered property itself and lacks legal standing, it cannot own property or enter into contracts to transfer patent rights to its human owner. In the DABUS cases, the human creator of the AI argued that he should own the patents simply because he owned the machine, a concept similar to owning the fruit of a tree. Courts rejected this, ruling that someone cannot be assigned patent rights from an entity that had no legal capacity to hold those rights in the first place.

Ultimately, patent offices have made it clear that their role is to apply existing laws, not rewrite them. While AI continues to play a massive role in modern research and development, current global statutes dictate that human involvement is mandatory to secure a utility patent. Until legislatures around the world actively vote to amend their intellectual property laws to account for machine-generated inventions, AI will remain legally classified as a tool rather than an inventor.