On 30 October 2019 the USPTO announced that it was seeking public comment on the possible impact of artificial intelligence (AI) on trademarks, copyright and other forms of intellectual property. With respect to trademarks, two specific questions were asked:
Would the use of AI in trademark searching affect the registrability of trademarks? If so, how?How, if at all, does AI impact trademark law? Is the existing statutory language in the Lanham Act adequate to address the use of AI in the marketplace?
The office has now published the 100 submissions received, scrutiny of which provides insight into how the trademark, creative and technology industries expect AI to change legal and office practice. Below are some of the key predictions, warnings and questions.
AI can help with search and risk assessments
A number of submissions noted that the use of AI in trademark searching can have a positive impact by improving and streamlining the trademark search and registration process, the APP Association noting: “Algorithms can handle tasks in seconds that would take trademark professionals and USPTO staff many hours to accomplish, saving time and reducing costs.” Going further, the American Bar Association suggested that, for trademark owners, AI could be used to more objectively evaluate the risks of adoption of a mark and in particular the risks of refusal by the USPTO based on a likelihood of confusion with one or more prior registrations. In reality, this is already happening.
AI could result in lower application numbers
Such use could affect examiner workloads. One suggestion by the American Bar Association was that the adoption of AI could lead to an overall reduction in the number of applications filed, and consequently, reduce the burden on the USPTO examination division. For instance, in helping examiners “find potentially fraudulent or digitally altered use specimens through evaluation of meta-data and comparison of images on a pixel-by-pixel basis, which examining attorneys may not be readily able to discern with the naked eye or without the benefit of having reviewed other cases in which the same or similar use specimens were submitted”.
It could also lead to more refusals
While the use of AI may guide applicants to lower risk filings, there is another potential impact according to the American Intellectual Property Law Association (AIPLA): “AI searching might enable USPTO examining attorneys to find more potentially similar design marks than the current design code indexing method does, and this could lead to more refusals to register on the basis of section 2(d).”
AI could be a tool for infringers
While AI could be an important aid for rights holders, it could equally prove to be a boon for infringers. Reflecting on its studies in criminology, The Center for Anti-Counterfeiting and Product Protection (A-CAPP Center) at Michigan State University noted that, once a criminal or civil infringer has access to AI, they can create algorithms that do exactly the opposite of the brands， which can “allow the infringer the opportunity to train their algorithm to initiate multiple postings with modifications to language, images or other key features with a speed and efficiency that individuals cannot match. AI can be used to actively scour platforms for brands and products, get into regulatory agencies’ websites, search for high sellers, run them against the database, see what’s protected and what’s not, and in theory choose other jurisdictions in which to register.”
How AI will affect trademark law
Actually, the jury remains out on this question. While some felt that “treating an AI as a tool used by a user, and the output of an AI as a ‘work for hire’ should generally fall in line with existing trademark law” or that the courts should develop case law as disputes arise, others saw potential issues. For example, the A-CAPP Center noted that ‘intent’ of the defendant is an important element in question when looking at the Lanham Act and how AI may interact with it, adding that in the case of using a counterfeit mark, the defendant must have “intentionally use[d] the mark or designation, knowing such mark or designation is a counterfeit mark”. This raises questions such as whether AI can intend for something to happen: “Is the onus on the user of the AI or the creator of the AI to not ‘intend’ for a counterfeit mark to be used?”
IP offices are already seeking to unleash AI
INTA’s submission revealed data from its survey of IP registries, which saw nine (Australia, Chile, China, Japan, Norway, Russia, Singapore, the EUIPO and the United States) provide insight into their use of AI. Many are already in the early stages of adopting AI solutions, with five developing trademark image search systems that incorporate AI. INTA notes: “Registries in the developmental stages were hopeful that the incorporation of AI in their search systems would enhance search processes…” The clear trend is towards positive engagement with AI solutions in the IP office environment.
The future is not yet here but it is fast approaching and the utilisation of AI by applicants and offices is set to significantly impact practice.