In the Public Eye: USPTO Issues Report on AI

The USPTO recently released the report “Public Views on Artificial Intelligence and Intellectual Property Policy”. The report is part of the USPTO’s effort to engage with the innovation community and experts on AI and to promote innovation of AI through appropriate intellectual property incentives.

The report includes the analysis of nearly 200 responses received from individuals and organizations to federal notices published in August and October 2019 to solicit public comments on patenting AI inventions and the impact of AI on other areas of intellectual property policy. The USPTO requested feedback on issues such as whether current laws and regulations regarding patent inventorship and authorship of copyrighted work should be revised to take into account contributions other than by natural persons.

AI in Evolution. As an initial matter, commenters noted that AI has no universally recognized definition, and any definition used as part of an AI policy must be dynamic enough to evolve as AI technology evolves. Some suggested that the USPTO revisit the question of non-human inventions when artificial general intelligence (AGI)—AI that mimics human intelligence—is a reality and not just “purely hypothetical.”

Sufficient and Not Necessary. The majority of respondents took the view that current U.S. IP laws provide sufficient protection for development using current AI technology. To many, existing contract law principles can be used to adequately fill in any gaps as AI technology further advances. Generally, commentators were divided on the need for new intellectual property rights to address AI inventions. Those focusing on new protections were focused mostly on data, with some suggesting that advances in AI should warrant more protection for data rights, including sui generis protection.

Human Not Machine Inventors. With respect to patents, commenters agreed in large part that, for now, humans, not machines, must be inventors. Further, most agreed that only a natural person or a company, through an assignment, should be considered the owner of a patent or an invention, although some suggested extending ownership to those who train an AI process or own or control an AI system. Other respondents were concerned on the practical effects of recognizing non-natural inventors, e.g., how would a machine sign an oath?

Impact on Patentability. Many commenters thought that AI-related inventions were not different than other computer-implemented inventions that are patent-eligible. Commenters were divided, however, on whether the growing presence of AI would affect the legal hypothetical standard of a “person having ordinary skill in the art”, i.e. the reference used to evaluate whether an invention is obvious. Another focus was the need for adequate disclosure and the unique challenges given the “black box” nature of certain AI algorithms that do not produce predictable results.

No Right to Copyright. The request for comment asked whether a work produced by an AI algorithm or process, without the involvement of a natural person contributing expression to the resulting work, should qualify as a work of authorship protectable under U.S. copyright law. Most said no, with one respondent noting that AI is a “tool, similar to other tools that have been used in the past to create works.” A minority suggested that creative works made by AI without human authorship should still be copyrightable—with the owner/controller of the AI system or the person who fixes the work in its final form as author. To most, authorship required human creativity.

Paying for Ingestion. The report also tackles the issue of mass digitization and text and data mining (TDM), noting that it “may be considered copyright infringement or fair use, depending on the facts and circumstances at issue.” While commentators considered the doctrine of fair use to be flexible enough to address the issue, some commentators suggested new mechanisms should be created to address licensing and compensation for use of digitized content for purposes of machine learning (ML) and ingestion.

Likelihood of Confusion. The request asked for comments on whether AI will impact trademark law. In this area, commentators were less animated. Most said that either the use of AI software would not impact trademark law or that the existing legal framework for trademarks is adequate. But some did comment that the effect of bias in algorithms should be accounted for if used in the registration process and that voice-activated applications to assist with registration could lead to a focus on phonetic similarity between marks.

Trade Secret Protection or Not. The report also touches on the importance of trade secret law to protect AI innovations and big data. Commentators acknowledged the limitations of trade secret law where collaboration and innovation require data sharing and where the use of AI could enable the discovery of information historically maintained in secrecy. Even so, commentators were not clamoring for changes to trade secret laws.

The report highlights many of the concerns raised around AI and IP rights as the technology of AI evolves. The USPTO intends to use the report to direct its focus on specific issues to further explore in connection with IP rights for related AI. Expect to hear more as AI continues to evolve.