The World Intellectual Property Office (WIPO) held its third “Conversation on Intellectual Property and Artificial Intelligence” on November 4, 2020, to discuss its revised issues paper on Intellectual Property Policy and Artificial Intelligence. This session identified the following issues:
- Defining artificial intelligence (AI) and future-proofing its definition as the technology evolves, exploring what is AI inputted and AI generated.
- The impact of AI on trademarks and the implications of human perception to determine registration and infringement of trademarks.
- The role of intellectual property (IP) policy in bridging the capacity gap.
- The policy implications of using AI in IP administration.
Public bodies in the United States, United Kingdom, and European Union have each recently published reports on the interrelationship of AI on IP policy. In October 2020, the United States Patent and Trademark Office (USPTO) published a report, Public Views on Artificial Intelligence and Intellectual Property Policy, on two formal requests for comments, and the European Parliament published a report on intellectual property rights for the development of AI technologies. In September 2020, the UK’s Intellectual Property Office (UKIPO) published a call for views on the policy considerations and future relationship between AI and IP.
The reports share common concerns with the WIPO: (1) defining AI in order to be technology neutral and flexible to future developments; (2) AI and related technologies that are based on computational, mathematical methods may not be patentable per se; (3) AI-generated creations pose regulatory challenges beyond those of AI-assisted creations; and (4) copyright infringement and privacy concerns around the use of data by AI. Courts in each jurisdiction have so far rejected the suggestion that AI has its own legal personality.
A majority of respondents in the USPTO’s report considered the current US IP statutory framework and legislation to be "calibrated correctly to address the evolution of AI" and that existing commercial law principles may fill any gaps left by IP law as AI technology evolves. Respondents classified AI as a subset of computer-implemented inventions and found that current USPTO guidance is equipped to handle advancements. A common theme was that AI is not sufficiently advanced—and will not be for some time—to warrant the exclusion of a human inventor, therefore eliminating the need for novel legislation. However, the growing ubiquity of AI would affect how the USPTO and courts would assess the legal hypothetical standard of a “person having ordinary skill in the art,” which is critical to determining the issuance of a patent right.
Many respondents felt that AI has the potential to affect the threshold for the US IP concept of “nonobviousness” by altering the skill level of the hypothetical “ordinary skilled artisan.” An invention that would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention is not patentable.
As to copyright, most respondents agreed that the current requirement for a human author (with specific exceptions) should remain.
The USPTO also sought opinions on the adequacy of the Lanham Act to the protection against trademark infringement by AI technology, due to the lack of personhood attributed to AI. Overall, respondents felt that the existing statutory and common law framework for trademarks in the United States is "sufficiently flexible" to address AI software.
The USPTO will further explore measures it may take to bolster understanding and reliability of IP for emerging technologies such as AI. It is also working to ensure that appropriate IP incentives are in place to encourage further innovation in and around this critical area.
The European Parliament (EP) appears to take a more proactive approach to the interrelationship between AI and IP. The EP’s report builds on the European Commission’s White Paper on AI policy published in February 2020, and emphasizes the importance of creating an “operational and fully harmonized regulatory framework in the field of AI technologies.” In contrast to the United States, only 3% of respondents to the European Commission’s White Paper thought that the current legislation is fully sufficient. The EP suggests that any future regulatory framework should take the form of a regulation, rather than a directive, to harmonize laws across the European digital single market.
The EP, like the USPTO, emphasizes that the main functions of intellectual property rights are to protect the interest of human creators and encourage innovation. The EP agrees that certain AI-generated works may be construed as “intellectual works” and be capable of copyright with the ownership assigned to an ultimate human creator.
The European Commission is expected to publish a draft legislative proposal on AI in early 2021.
The UKIPO’s call for views explores patents, copyright, design rights, trademarks, and trade secrets. It asks whether AI systems can devise inventions and whether there is a moral case for recognizing AI as an inventor of a patent. The UKIPO also asks if greater clarity is needed about who is liable when AI infringes a copyright, and whether content generated by AI can be eligible for protection by copyright or related rights.
The UKIPO’s consultation is open until November 30, 2020, and its report is expected to be published in early 2021.
At the WIPO’s session on November 4, 2020, the UK Minister for Science, Research and Innovation called for a shared definition of AI to be agreed and used globally.
The WIPO and each jurisdiction above have identified issues and flexibility in the current law, but no policy or legislative solutions as yet. The United States appears to be broadly adopting a “wait and see” approach to IP policy while AI technology evolves. The United Kingdom awaits public feedback and is seeking to build an international and consensual approach to IP policy and AI, beginning with a shared definition of AI. The European Union has given the strongest indication of the three jurisdictions of legislative action to address the interrelationship between IP and AI.
Courts may force idiosyncratic solutions if public bodies do not identify IP policy solutions quickly enough. The WIPO is expected to publish a White Paper in 2021 that may propose a definition of AI for the purposes of IP policy, which could, if agreed by WIPO member states, trickle down to national IP bodies’ approaches and find its way into legislation.