Evolution of the Law of Trademarks and its processes by the use of AI

Contributor: Sneha Dutta

“AI is a powerful river, flowing swiftly to boost productivity, yet requiring careful navigation to avoid its hidden currents.”

AI has been a boon in a number of ways for the intellectual property sector. From effectively helping to catch hold of the infringements by identifying the deceptive similarities amongst the existing brands, to conducting speedy and accurate trademark searches, AI has increased the efficiency of the entire trademark management system manifold. The processes involved in trademark management system have become quicker and speedier like never before. But, the evolution and the changes introduced in the law of trademarks by AI is not all roses. It comes with legal repercussions and challenges to be dealt with. While trademark is all about the interaction between the consumer and the brands, there is a considerable dilution in this relationship pursuant to AI’s role as a filter between the two. This influences the share of purchasing decision and autonomy of the consumer which in turn disturbs the equilibrium of the fundamental functionality of the law of trademark. Moreover, in cases of AI-generated trademarks of brands, several questions remain undetermined. There exists an unresolved issue pertaining to the ownership of the trademark between the AI-developer and the one who inputs the prompts for the AI-based brand generation. Additionally, the ascertainment of the entity that is to be held accountable during trademark infringement process involving an AI-generated trademark is difficult.

Role of AI in boosting the efficiency of trademark processes

AI trademark search algorithms are trained on vast datasets of existing trademarks to identify patterns and even the minute similarities. By combining advanced image recognition, Natural Language Processing (NLP), and Machine Learning (ML) techniques, AI systems are performing comprehensive trademark searches and flag potentially conflicting marks. As a result, it has been possible to lessen the cases of infringement of the already existing registered trademarks. Machine learning models are trained and made familiar with historical trademark data in the database so as to use them in predictive analysis of the chances of acceptance or rejection of a trademark received for registration. 

  • Computer Vision technique is trained in reading through the textual data on images and hence is employed in detecting visual similarities of trademarks as well as on trademark logos. 
  • Models of Predictive analytics help in the anticipation of future trademark conflicts by using the information of past trademark data and market trends. This facilitates management as well as entrepreneurs to make better decisions for their brands in respect of their trademarks. 
  • AI technologies like Natural Language Understanding (NLU) and Robotic Process Automation (RPA) are used to automate the repetitive time-consuming tasks involved workflows of trademark registration thereby enhancing the productivity levels and bringing an overall efficiency.
  • Intelligent Recommendation Systems are a guiding tool to the trademark applicants to assist in doing away with errors in the trademark to get it registered successfully. Hence, with AI, a chance to protect the mark from being rejected is thereby unlocked. 

AI can proactively scan various platforms like e-commerce websites and social media in order to detect any real-time trademark infringement by way illegal usage in the form of counterfeit goods. The real-time brand monitoring helps to initiate prompt actions to guard against brand dilution and preserving the brand identity. 

Due to AI, it has been possible to automate the trademark enforcement process. The automation of the enforcement procedure has resulted in proper utilisation of resources and quicker resolution of trademark disputes. While AI is capable of carrying out the repetitive steps of issuing cease and desist letters, takedown requests in an automated fashion, it is probable for the disputes to remain pending when handled manually. This is because AI is trained to analyse vast amount of data in databases quickly and therefore, it can execute the enforcement procedures as per the stage of each case. But, on the other hand, a higher possibility exists for humans to miss out on deadlines of the applications to be made even if a manual database is maintained.

Legislative changes required to be brought in the existing trademark laws for an effective incorporation of AI in the trademark processes

The aspect of goodwill and reputation concerning the trademarks stems from the emotional connection that the consumers develop with the quality of the product. But, with the involvement of AI that analyses brand data by the use of algorithms, this concept might become obsolete due to the deduction of the human elements that form the very base of it. 

The traditional concept of likelihood of confusion in the minds of the consumers might also fade away in the era where the law of trademarks meets AI. The accuracy that AI holds in detecting the intricate details and minute similarities by scanning through huge pool of data makes the chances of missing out on similar brands very bleak. This occurrence is opposed to how consumers perceive likeliness between brands. Pursuant to this, there is a need for this concept to take a shift as confusion on the part of AI is unlikely.

What constitutes infringement as per section 29 of the Trademarks Act, 1999 is entirely based on how human mind perceives the similarity and identical nature of markets. But, with the incorporation of AI into the system, this provision would need necessary changes since the possibility of likelihood of confusion between marks has become considerably lesser. This is due to the ability of AI to detect even the minute resemblances unlike human mind. Moreover, as per the predictions made by the AI tools, within the next five years, the text search is going to be replaced by voice search as 30-50% of the product searches would be conducted by voice search by then. That being said, the importance of visual similarity might gradually turn blurry with a greater significance of phonetic similarity in place.

Case Laws on AI and trademark infringement

In the cases of L’Oréal v eBay and Louis Vuitton v Google France, it has been held that the AI application provider would not be liable if it was unaware of the infringing activity and had taken adequate steps to take them down. In the aforementioned cases, Google and eBay were not held liable respectively as it was found that they had not joined hands in conducting the infringement. 

But, in the case of Cosmetic Warriors Ltd and Lush Ltd v Amazon.co.uk Ltd and Amazon EU Sarl, Amazon was held liable for trademark infringement because Amazon’s use of AI algorithms led to the display of the trademark ‘LUSH’ on google ads and redirected to Amazon website even when Lush products were not sold on amazon. This affected the brand reputation of the petitioner and caused confusion about the origin of the Lush products that were advertised. Hence, it was held that if the company using AI algorithms is a part of the infringement, then, it would be held liable.

Conclusion

AI tools know no geographical boundaries or linguistic barriers to function and hence, has the capacity to provide for a robust protection strategy to the trademark owners. AI has immense potential in combating infringement by providing for innovative measures and tools for an effective screening to catch hold of infringements. But, the ethical and legal considerations including concerns around data protection and privacy, come along with the benefits it has to offer. In order to leverage AI in the trademark processes as a resultant advantage, it is necessary to inculcate AI into the system with a thoughtful and chalked out approach to address the associated challenges. Moreover, a complete automation would also not lead to an optimization as human involvement is significant in decoding context. Hence, a balance has to be achieved between automation of the trademark processes by AI and human oversight.

Scroll to Top