In the age of Artificial Intelligence (“AI”) proliferating, disruptions are experienced across industries including service and media. Generative AI (or “GenAI”) is also interluding in the realm of intellectual property (“IP”). Trademark laws, among other IP Laws, are facing new challenges and complexities and require fresher perspectives for improving brand protection, especially in countries like Singapore, holding as a tech, information technology and IP hub.
Understanding Generative AI and Trademark Law Challenges
Gen AI is prevalent on both the supply and demand ends of commerce. New brand owners are likely to use generative AI to conceptualise trademarks while studies have also shown consumers, especially the younger demographic consisting of the Millenials and GenZers, are more likely to be persuaded by recommendations made by AI, such as ChatBots, in their consumption patterns. One such introduction of ChatBots with built-in AI is by the ASEAN e-commerce giant Lazada, under the name “Lazzie”. These ChatBots recommend suitable purchases based on algorithms and machine learning of the youngsters’ browsing and consumption history, amongst other things.
From the trademarks’ perspective, some important considerations include:
(a) Who owns the mark?
(b) Does the trademark generated through AI pass the traditional tests of registrability?
(c) Are decisions made by an “average consumer” organic anymore? Would the decisions made through AI be considered an “informed decision”?
(d) Would an AI-generated mark be clear of legal disputes including but not limited to infringement and passing off?
The analyses of each of the above questions are comprehensively dealt with in the recent publication of the Intellectual Property Office of Singapore, a brief summary is as follows:
Ownership of trademark when generated by an AI is quite literally a tug-of-war between the user who leads the AI with questions and the AI that optimizes its cognizance. While Singapore’s IP laws indicate a mark’s ownership to be register dependent, plainly, this needs more scrutiny and is not definitive yet.
Traditionally, a mark’s registrability tests are devised around its capability of being “unique”. With certain bases for assessment such as visual/aural/conceptual similarity, distinctiveness, examiners have been gauging its uniqueness and thereafter, granting the mark protection. At this stage, GenAI’s ability to conceive a mark that is unique, has yet been tested.
Trademark laws are premised upon the doctrine of first impression and imperfect recollection of an “average consumer” who is a human. As mentioned earlier, consumers in this day and age are persuaded by AI’s recommendations, which have the capability to precisely distinguish marks products/services from each other. The decision of the purchaser is reliant on the information presented by the AI.
Therefore, the average consumer is no longer the target for brand protection, which takes a toll on the underscores of trademark laws. The examination of marks now has to elevate to the next level, and considerations of a mark’s uniqueness should be looked at from the AI’s perspective.
Infringement and passing off concerns may also be triggered when GenAI is conceptualising the mark, mainly because it builds on available datasets and the suggestions carry the potential to infringe IP rights of existing brand owners. Therefore, brand owners should be mindful of this very concern.
While the legal challenges are plenty, integrating AI into the IP ecosystem is the duel of equals that is to be expected. For now, to preclude the potential problems that brand owners using GenAI tools might face, pre-filing clearance searches are encouraged. This promotes good due diligence practice, and assessment from a standpoint of registrability.