WIPO Revises and Expands AI Policy Issues

The World Intellectual Property Organization (WIPO) has published a Revised Issue Paper on Artificial Intelligence and IP Policy. The Revised Issues Paper updates the Draft Issues Paper published in December 2019 with the addition of a glossary and sections on trademarks and trade secrets, and an expansion of the sections covering patents, copyright, infringement and deep fakes.

As reported by IPWatchdog in February, WIPO received more than 250 submissions from IP offices, companies, organizations and individuals in response to the draft paper. These have been taken into account in the Revised Issues Paper, which lists a large number of questions (many of them not included in the draft paper) under 16 separate issues.

WIPO will host the second session of the WIPO Conversation on IP and AI virtually on July 7-9, 2020 daily from 1.00pm to 300pm CET. The Provisional Agenda is based on the Revised Issues Paper and registration can be made online.

Definitions Provided

Following criticism in some of the submissions that the draft paper did not define terms such as AI and “autonomously generated by AI,” the Revised Issues Paper includes a list of working definitions. It defines AI as “a discipline of computer science that is aimed at developing machines and systems that can carry out tasks considered to require human intelligence, with limited or no human intervention” and says that the paper equates AI to “narrow AI” namely “techniques and applications programmed to perform individual tasks.”

It adds that “AI-generated” and “generated autonomously by AI” are used interchangeably to refer to the generation of an output by AI without human intervention, and are distinguished from “AI-assisted.”

The glossary (issue 1) additionally defines the terms “output,” “literary and artistic works,” and “data subsisting in copyright works.” However, it also raises three questions about where the line should be drawn between AI-generated and AI-assisted; further terms that need defining; and how to define terms in a technologically neutral way.

Inventorship, Ownership and Patents

Many of the submissions made to WIPO addressed questions about AI-generated inventions, such as those raised by the decisions on the Dabus patent applications by the UK IPO, EPO and USPTO. The patent sections of the Revised Draft Paper reflect these concerns.

Under Issue 2, the Revised Issues Paper states that “the role of AI in the invention process is increasing,” which compares with the statement that “inventions can be autonomously generated by AI” in the draft paper. It identifies a total of eight questions relating to “AI-generated inventions,” including whether the law should permit an AI application to be named as an inventor or, if a human inventor must be named, how that should be determined; and who should be recorded as the owners of a patent involving an AI application. It also raises the questions whether AI-generated inventions should be excluded from patent protection, and the impact on infringement, liability and dispute resolution.

The Revised Issues Paper provides more context and sets out five questions on patentable subject matter and patentability guidelines (issue 3) and inventive step or non-obviousness (issue 4). On the latter point, it has added a new question asking, “is it necessary to retain the traditional requirements of inventive step or non-obviousness, which are fundamentally associated with human acts of invention?”

On disclosure (issue 5), another new question has been added: “How do the current rules on disclosure apply to AI-generated and AI-assisted inventions and are they sufficient to meet the underlying policy rationale?” Issue 6 covers general patent policy considerations and asks specifically whether AI-generated inventions should benefit from patent protection and whether consideration be given to a sui generis system of IP rights?

Other IP Rights

The changes regarding copyright and related rights, data and designs (issues 7 to 12) are not substantial, although there is a new question: “Do AI-generated works require copyright or a similar incentive system at all?” And a few other questions have been added, notably on deep fakes and designs.

Two new sections cover trademarks (issue 13) and trade secrets (issue 14). The Revised Issues Paper notes that “there may be areas of trademark law that may be impacted by AI” and states that “questions arise with respect to registrability, infringement and also unfair competition.” It sets out six specific questions.

On trade secrets, the paper states that “trade secrets are increasingly relied on to protect investment in collecting and curating data and innovations in the AI field” and identifies seven questions covering the balance between protecting innovation and third parties’ interests, whether data and AI applications should be protectable by trade secrets or other IP rights and the need for harmonization.

Issue 15 covers capacity building and includes a new question: “What kinds of cooperation mechanisms between countries of similar technological development in AI and IP could be envisioned?” Finally, issue 16 is accountability for decisions taken in the prosecution and administration of IP applications, covering the legal questions raised by the use of AI applications in decision making, the types of decisions that can be taken by AI applications and appeal mechanisms.

A Rich Source for Discussion

The revised questions are likely to provide a rich source for discussion at the virtual Conversation on July 7 to 9, and anyone interested in this topic is encouraged to consider taking part. The provisional agenda includes presentations from H.E. Mr. Oman Sultan Al Olama, Minister of State for Artificial Intelligence (U.A.E.); Mr. Ahmed Elgammal, Professor, Department of Computer Science, Rutgers University (U.S.); Judge Klaus Grabinski, Federal Court of Justice (Germany); and Judge Kathleen O’Malley, U.S. Court of Appeals for the Federal Circuit.

Image Source: Deposit Photos
Author: sdecoret
Image ID: 212545556