We’re pleased to bring to you a guest post by Shubham Borkar and Nitish Daniel. Shubham is a Senior Associate at Khurana and Khurana Advocates and IP Attorneys and Nitish is an Assistant Legal Advisor at ONGC.
Need for Boosting Trademark Law for Keeping Pace with Artificial Intelligence
Shubham Borkar & Nitish Daniel
Business Giants like Google, Facebook, IBM are integrating AI systems into their operations, so are online retailers, online marketers, and product manufacturers like Amazon. For example, you upload a picture on Facebook, it suggests you the name of the person you want to tag. Amazon Inc. admits that without machine learning based AI systems, it won’t be able to grow its business, improve its customer experience and selection, and optimize its logistic speed and quality. Amazon recently announced a new initiative called ‘Project Zero’ that will use AI to detect and remove counterfeit products. This initiative is in its initial stages, but eventually is likely to be implemented globally. So, marketing and product branding is being affected by AI revolution. In this post, we will be discussing the implications likely to arise in trademark law due to AI.
What’s Artificial Intelligence (AI)?
The definition of artificial intelligence (AI) hasn’t been stagnant, rather ever involving. The term ‘artificial intelligence’ was first coined by John McCarthy in 1956, the focus of which was ‘thinking machines’. Eventually, the meaning evolved with the modern meaning focusing more on the ability of the machine to imitate human intelligence. According to Oxford English Dictionary, ‘artificial intelligence’ is “The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.” Frankly, it can be said there are too many definitions of AI, but this simplistic definition gives us enough understanding to analyse the legal implications surrounding the field. AI systems can be classified into different categories, the main one being strong AI (systems think and perform tasks exactly like a humans) and weak AI (systems focusing just on one narrow task).
We might feel that trademark law has always been the same and has always served the same purpose, but the way we shop has undergone multiple revolutions and each of these revolutions have influenced several changes in trademark law. Trademark law was laid down in the era when shopkeepers or dedicated shop assistants used to suggest the products to consumers which the consumers bought. As these shopkeepers or shop assistants were well versed with every product’s details and trademarks, the likelihood of them getting confused with other deceptively similar marks was infinitesimal. Then came the era of self-service stores in 1916 with the Piggly Wiggly chain of stores of Memphis USA which revolutionized the way we buy products. With the advent of this era, the consumers made shopping choices themselves based on the reputation of the brand. Consequentially, the value trademarks served increased multiple folds, but so did the likelihood of consumers being confused between trademarks. In the 1990s we were hit with another revolution of online retailing with which trademark law faced the challenges of incorporating new concepts of domain names, metadata, meta-tags, initial interest confusion etc. The social media revolution of the 2000s impacted our buying choices by suggesting products while we scroll the social media feeds.
Now, with the advance of AI, another revolution seems to be taking place in the consumer market. These online retailers are using AI to recommend us products that we would like to buy by analysing our buying profile, search history and many other things. Trademarks served as a means to help consumers choose a product they trust and want to buy. AI in the future is likely to deprive the consumer of this choice and remove the consumer from the decision making process as to what product to buy. Analysis by Mr. Ajay Agrawal, author of the book ‘Prediction Machines: The Simple Economics of Artificial Intelligence’ reveals that AI systems will reduce the cost of accurately predicting an outcome drastically. So, he predicts that in the future, marketing strategy would change from ‘shopping to shipping’ to ‘shipping to shopping’. Online retailers would be able to accurately predict your needs using AI and then ship out the product to you, you will try the product and if you like it, the payment will be deducted. They will be able to do so as AI outcome prediction would be extremely cheap, enabling them to invest more into efficient shipping technologies like drones as well as product return infrastructure.
Currently, we are seeing a surge of products like Amazon Echo, Google Home, and Apple HomePod that provide AI assistance to humans. Consider a scenario where such an AI assistant is asked by a person to buy a product based upon the predetermined standards for buying that product such as quality and quantity. Such ‘Amazon Echo’, ‘Apple HomePod’, or ‘Google Home’ orders a counterfeit/infringing product, would then Amazon or Google be liable for secondary infringement along with the infringing product manufacturer. How will the standards of ‘likelihood of confusion’, ‘imperfect recollection’, and ‘average consumer’ applies to AI? Or should they at all be applied to the AI?
Same issues will have to be considered in case of another emerging revolutionary technology called the Amazon Dash Replenishment Service (DRS) which enables connected devices to order physical goods automatically from Amazon when supplies are running low. The customer selects the products they want to automatically reorder and the device supporting DRS measures and tracks usage and when the supplies are running low, places an order using DRS, and Amazon then ships the product to the customer. In future such service may have the discretion to choose the product as well.
If this seems too far fetched to you then there are AI chatbots in the marketplace that act as personalized shopping assistants, an example would be eBay’s ShopBot, H&M’s Kik. Surveys reveal that consumers are comfortable with sharing their personal information to receive better recommendation on products and a good AI chatbot influences their loyalty. What if ShopBot suggests an infringing product to the consumer and the consumer buys it? Shouldn’t such AI assistants be forced to be indiscriminate between brands while suggesting a product, otherwise it will keep recommending those brands which pay such AI assistant corporation?
The beauty brand Coty has partnered with Amazon to launch a new skill for the Echo Show, which is the first Alexa-enabled voice device with a screen. The interface allows users to input personal attributes, including hair, eye and skin color, and Coty supplies an on-demand, occasion-based look planning service, capable of delivering more than 2,000 combinations of hair and makeup looks. Along with curated tutorials and quick tips, thereby acting as a user’s personalized stylist. The consumer will be able to choose the look they like and order the products required to achieve that look via Amazon Echo voice. Shouldn’t such AI assistants be forced to be indiscriminate between brands while suggesting a product, otherwise it will keep recommending those brands which pay such AI assistant corporation?
All this demands revision of concepts like ‘imperfect recollection’, ‘likelihood of confusion’, ‘the average consumer’, and ‘secondary infringement’ etc. which form the foundation of trademark law, as all these are centered around human beings and their capabilities. In Cadila Healthcare Limited vs. Cadila Pharmaceuticals Limited,[1] Supreme Court of India clarified that average consumer is the one with average intelligence and imperfect recollection. AI will have neither average intelligence nor imperfect recollection and there would be less likelihood of confusion, so perhaps trademarks could be closer looking? How would the courts apply the ‘average consumer’ standard to AI? Would these online retailers be liable for secondary infringement?
AI is also likely to influence how trademark enforcement will be done. According to Charlie Hill, head of product at TrademarkNow a global IP resource believes that because of factors like numerous applications and registrations of new trademarks, unsynchronized and disconnected trademark offices, and imperfect analysis of trademarks by human eyes, humans aren’t well suited for trademark analysis and we must incorporate AI based trademark searching and monitoring services. As mentioned above, Amazon’s Project Zero is an anti-counterfeiting measure introduced by Amazon that uses AI. So, in future more trademark enforcement measures are likely to be introduced that will use AI.
In conclusion, with the current AI revolution that is taking place, every legal field is going to be affected and trademark law would also not be spared. Therefore, we need to constantly keep exploring how it will affect trademark law and try to keep the law at pace with the AI development, so that these trademark issues don’t become loopholes. Currently, nobody can definitely suggest how solutions to these issues are likely to unfold. But, by being positive and influencing law the correct way, we can hope that the ultimate result is the elimination of counterfeiting, quality products for consumer, reduction of monopoly, and increase in competition.
