Artificial Intelligence and Trademark Law
The impact of artificial intelligence (“AI”) technologies on trademark law has been among the frequently debated topics over the past year. As also discussed in our article published last year, while these technologies accelerate processes in terms of trademark registration procedures, they simultaneously necessitate a re-assessment of the human-centered criteria traditionally applied in trademark law.
Indeed, Intellectual Property (“IP”) offices worldwide continue to increasingly adopt these technologies in order to enhance time and cost efficiency, improve the accuracy of procedures, and productivity. In this regard, the World Intellectual Property Organization (“WIPO”) maintains an Index[1] listing country-based examples of such use. These include AI-powered translation tools, systems assisting in the identification/classification of goods and services during trademark application processes, applications enabling word-based, semantic and visual similarity searches as well as reverse image searches, and self-learning systems that continuously improve over time.[2] In Türkiye, the Turkish Patent and Trademark Office (“TÜRKPATENT”), within the scope of its “AI Supported Digital Transformation Project”[3], aims to integrate application processes, streamline workflows, and accelerate examination periods through a new AI-supported software infrastructure.
Another issue that should be addressed is that AI technologies necessitate a re-evaluation of the general principles of trademark law. Indeed, with the rise of AI technologies, situations have begun to emerge in which consumers’ purchasing preferences are shaped by the results presented to them by AI, or in which the act of purchasing itself is carried out by AI directly upon the consumer’s voice instructions. Looking ahead, scenarios in which such transactions are automatically performed by AI without the need for explicit instructions have also sparked debate.
Traditionally, trademarks have functioned as tools that simplify complex product information and provide a reliable indicator of origin to serve consumers’ limited cognitive processing capacity. However, in scenarios where purchasing processes are conducted by AI systems, it is considered that algorithms would simultaneously analyse multiple data such as price, quality, user reviews, supply chain data, and personal preferences, owing to their capacity to process large datasets. In such cases, trademarks may cease to be the primary determinant in the decision-making process and instead become merely one of the data points evaluated.[4] From this perspective, for a computer, there may be no meaningful distinction between a trademark and a line of code. However, in our view, the emotional meaning attributed to trademarks through the human experience, as well as the significant impact of the communication function of trademarks, continue to be crucial. Accordingly, even as AI assistants and recommendation algorithms become increasingly widespread, the possibility that decision-making processes will be entirely delegated to AI systems without human interaction should be assessed with due consideration of these factors.
Another issue is that the current system is based on the perception of the human-centered ‘average consumer’ and relies on concepts such as the ‘likelihood of confusion’, which are assessed through visual, phonetic or conceptual similarity, taking into account human traits such as limited attention, imperfect memory, and the tendency to inadequately evaluate context in the pace of everyday purchasing decisions. However, in situations where AI systems come into play, and considering their ability to analyse context in a holistic manner, it will no longer be possible to speak of confusion in the same sense as it applies to humans.[5] Instead, the likelihood of confusion or even error may arise due to factors such as algorithmic manipulation or the misguidance of data flows. There is no doubt that this will bring the aspects of data protection and competition law to the forefront when undertaking the issues.
In conclusion, within this evolving and transforming landscape, the increased use of digitalisation and AI systems may require a fundamental re-examination of the core principles of trademark law. In such an environment, when shaping their commercial strategies, trademark owners may benefit from taking steps toward digitalisation by developing system-compatible and machine-readable versions of their trademarks[6]; securing these through trademark registrations and further integrating appropriate technological tools such as blockchain, the Internet of Things, and other relevant technologies into their business operations, while also adopting preventive measures to avoid situations that may give rise to infringement.
[1] https://www.wipo.int/about-ip/en/artificial_intelligence/search.jsp
[2] https://www.wipo.int/en/web/ai-tools-services#:~:text=WIPO%20has%20developed%20an%20AI,Patent%20Classification%20(IPC)%20schema.
[3] https://www.turkpatent.gov.tr/haberler/turkpatent-akademi-tanitimi-ve-yapay-zeka-destekli-dijital-donusum-projesi
[4] Michael Grynberg, Kentucky Law Journal, AI and the Death of Trademark, Cilt 108, 2019-2020, number 2, page 205-2010; Alpana Roy . Althaf Marsoof, Removing the Human from Trademark Law, Springer, 22 April 2024, page 740-741.
[5] Roy- Marsoof, Removing the Human from Trademark Law, page 740-741.
[6] Roy- Marsoof, Removing the Human from Trademark Law, page 757.
First published by Gün + Partners in Apr 03, 2025.