U.S. Patent laws usually have two objectives: 1) To disclose inventions for the benefit of mankind; and 2) To incentivize inventors and investors. Thus, patent protection serves a social benefit and a personal benefit.

Currently, thousands of patent applications are being filed in the U.S. Patent Office for inventions directed to AI, and despite the patent eligibility issues, patents are being issued on AI inventions by humans. However, these patents are distinct from the possibility of inventions by computers using AI. Whether such inventions can or should be protected by patents, raises many issues for these thinking machines.

An early AI pioneer, Stephen Thaler, developed the “Creativity Machine" in 1994. This machine is credited with an invention which is covered by US patent 5,852,815 issued in 1998, the first known patent on an AI-generated invention. The ‘815 patent lists Thaler as the sole inventor, though his Creativity Machine is credited with generating the invention. Another computer scientist, John Koza, designed the Invention Machine based on genetic programming and biological evolution. This machine made an invention covered by US patent 6,847,851 issued in 2005, listing 3 people as inventors.

Thus, inventions generated autonomously by computers using artificial intelligence are here. But how do the patent laws apply to such inventions?

The starting point is the first patent statute, 35 U.S.C. § 100, which defines “inventor” as the “individual” who invents or discovers the subject matter of the invention. While a human wrote the software code for the computer, that programmer did not invent the solution to the problem. Rather, the solution was generated by the AI. Thus, the computer may be the inventor, and not the programmer.


100 provides that "Whoever unquote invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.” Does “whoever” imply a person, and not a machine?

102 defines the conditions for patentability, particularly novelty, and states, “A person shall be entitled to a patent unless…" Thus, §102 seems to preclude a patent for invention created by AI. However, §103 describes the non-obviousness requirement for patentability, and provides, “patentability shall not be negated by the manner in which the invention was made." Thus, § 103 seems to be inclusive of AI generated inventions, and inconsistent with §102.

103 raises another dilemma for AI inventions, by requiring that the differences between the claimed invention and the prior art must be nonobvious “to a person having ordinary skill in the art…." Should or can an invention by a computer be tested or compared to a person skilled in the art? If AI provides a novel and useful solution to a problem, which no human had solved, is that solution per se non-obvious?

In view of the anticipated growth applications for AI, perhaps it is time for Congress to revisit the patent statutes.

The patentability of AI generated inventions has not been addressed by either Congress or the courts. China’s New Generation Artificial Intelligence Development Plan refers to AI Intellectual Property rights, so China appears to be ahead of the United States on such rights.