Artificial Intelligence and WIPOJune 15, 2020

On May 21, the World Intellectual Property Organization, a global forum for IP services, policy, information and cooperation, issued a revised paper on intellectual property policy and artificial intelligence (AI).  Acknowledging that AI is a general purpose technology that has widespread applications, WIPO held an initial “conversation” with interested parties in September, 2019.  WIPO then published an initial paper in December 2019 regarding issues of AI and IP, and received more than 250 comments from government and non-government entities, such as commercial actors, research institutions, universities, professional organizations, and individuals.  Based on the numerous comments, the new revised paper focuses on the legal issues raised by AI and IP policy, including patents, trademarks, copyrights, trade secrets, designs, and data.  This brief article is limited to the patent issues of AI.

First, the paper defines “artificial intelligence” as a discipline of computer science directed to the development of machines and systems that can carry out tasks that are considered to require human intelligence, with little or no human intervention.  Thus, as a science, AI falls within in the goal of U.S. Constitution, Article I, section 8, clause 8, to promote the progress of science and the arts.  A fundamental objective of the patent system is to disclosure technological improvements so as to benefit the public by making the technology available.  The patent system also encourages investment of time and money to offset the risks of invention development which positively contributes to society. 

With the increasing role of AI in the invention process, and the filing of patent applications in Europe and the U.S which name AI as the sole inventor (see author’s prior article, Dear USPTO: Patents for Inventions by AI Must Be Allowed, published 5/21/20 by IPWatchdog), issues arise as to protection for such inventions.  First, do AI generated inventions need patent protection or a similar system of incentives?  If patentable, should the law require a human inventor?  If there is no human inventor, who owns such inventions?  Also, if there is no protection for AI inventions, will that lead to them to be protected by trade secrets, which decrease flow of information and technological improvements?

If there is to be protection of some sort for AI developments, should there be worldwide harmonization, for global consistency?  For example, computer software inventions currently may be patented in the United States, but not in Europe.

If AI inventions are patentable, what will be the test for inventive step/non-obviousness?  This requirement for patent protection is historically related to what would be obvious to a person skilled in the art.  How should this test change for AI inventions that have no human conception or reduction to practice?

WIPO plans to have another “conversation” in July to discuss these, and other issues.  As AI continues to evolve, reaching a worldwide consensus on how to protect non-human inventions will enhance not only the entities working in this technology, but also provide valuable and sometimes life-saving benefits to humankind. 

Kirk Hartung is a patent attorney and chair of the mechanical and electrical practice group at McKee, Voorhees & Sease, PLC.  For additional information please visit or contact Kirk directly via email at

← Return to Filewrapper

Stay in Touch

Receive the latest news and updates from us and our attorneys.

Sign Up