Consumption of goods is being made convenient and easier with the emergence of AI. For instance, AI is giving consumers recommendations on the various products they wish to consume by using a simple voice command (Siri, Alexa, Google Assistant) or through search recommendations (Steponénaité, 2019). The convenience that comes with AI cannot be downplayed; however, we also have to recognize the legal quandary that comes with these developments. However, for these dilemmas to be addressed, we need to first appreciate where we are legally to make strides. Therefore, to appreciate the impact of the law on trademarks, it is important to recognize how the process of purchasing goods has evolved, as this goes to the core of the laws on trademarks.
Evolution of purchasing of products
Over time, the manner in which people purchase goods has evolved and part of the evolution is attributed to the coming of the 4IR. There have been three forms of evolution in the process of purchasing products. First, in the Nineteenth Century, it began with having a salesperson at the shop informing the consumer of the available goods for purchase. Based on their recommendations, people would purchase certain goods (Curtis & Platts, 2017).
The process of purchasing products then evolved to the self-service mode of shopping. This is better explained by what happens in the supermarkets in Kenya. Various goods are displayed on the shelves and a consumer will enter the supermarket and pick what they want to purchase based on the information they have on the various brands. During this era, the law on trademarks was very effective because various manufacturers had unique marks on their products which helped to identify the goods. A consumer could purchase goods from a specific brand based on the quality associated with the brand and the goodwill established in the market; the essence of trademarks law. This period could be described as the hallmark of trademark law practice.
Technological advancements occurred and there came the rise of social media and the influencers. During this era, consumers were more likely to purchase goods that had been advertised by ‘influencers’ than exercising their judgement when purchasing goods (Curtis & Platts, 2020). As the modes of shopping evolved over time, e-commerce, commonly referred to as online shopping, also became a popular trend among people. With the rise of online shopping came the increased use of AI. There is no universally agreed definition of AI but in 1955, Professor John McCarthy defined AI as “the science of making intelligent machines”. It has been argued that AI is an attempt ‘by humankind to replace human beings’ existence. The machines are programmed in such a manner that they can perform tasks that are ordinarily done by humans but at a more efficient level.
How AI works is that the developer programs the machine by inputting certain information; in return, the machine will provide a specific desired result. This is the main foundation of e-commerce or online shopping. A good example of AI in e-commerce is Amazon Alexa, Consumer Chatbots, Jumia Shopping Platform, Uber eats, Glovo, and AI shopping assistants such as Mona. These are platforms and machines that people use daily for convenience. As AI is programmed to monitor and emulate human behavior in most aspects of life, it collects certain information from a person. It can either do that through the online searches made by a person or by listening into the person’s life and creating a niche based on the information gathered. Perhaps the perfect example of how
AI works is when you talk about getting yourself a new bag and then suddenly all your Instagram and Facebook feeds bring adverts of bags and this leans more towards targeted adverts that we shall not delve into here. In such a scenario, the AI infrastructure has identified a need from what you talked about and hence it is trying to make your life easier by giving recommendations.
Another example is when you want to buy a phone, and you search on your search engine for phones. Based on your browsing history, your social media platforms and the online-shopping platforms will recommend phones for you to purchase. Perhaps this is why it is important to note that the era of purchasing based on social media and influencers’ recommendations and AI are closely related.
The AI era is what the world is experiencing and it is predicted that there are more advancements to be experienced in the field of online-shopping in relation to the use of Artificial Intelligence including but not limited to self-replenishment AI infrastructure. For instance, Samsung’s Family Hub fridge is programmed in a manner that will automatically replenish your groceries that are about to be depleted (Samsung).
Nexus between AI and trademarks
Conversations around AI and Intellectual Property have revolved around copyright and patent; however, the effect of AI on trademarks has been downplayed, and this does not mean that AI has not impacted trademarks. How the traditional law on trademarks is shaped is that the mark will distinguish a person’s goods or services from those of another person. In most cases, the trademarks are associated with the quality of goods offered by a specific person or company. The quality of goods is an important aspect of trademarks law because there is a rise of counterfeit goods especially in the era of e-commerce and online-shopping.
Statistics show that most Gen Z’s prefer online shopping to physically purchasing a product from the shop (Gutierrez, 2022). When doing online shopping, various issues need to be taken into consideration. Firstly, the person will consider the price of the goods, the estimated amount of time before the goods are shipped and the availability of the product. For instance, when shopping on the Jumia online shopping platform, you will search for the goods you intend to purchase. After searching, you can sort the goods based on the price, discount, rating and shipping period. Once the goods have been sorted, the cheaper goods will likely come to the top and the expensive ones at the bottom. This means that the person shopping is not looking at the brand but rather the price; disregarding the tenets of trademarks law of associating quality with certain marks. Further, how AI is programmed is that it has good memory and if you have shopped for the same thing before, it is likely to bring you the same product from the same manufacturer as you did before or recommend similar products. At this point, one cannot tell if the recommendations are counterfeit products. Therefore, unlike in the past when people would visit a shop and look at all the brands that have put their goods on display and pick out their preferred brand, today, people purchase goods recommended by AI. The recommendations given by AI are based on the information made available to them and in most cases, it may not speak to the quality of a certain brand.
The tenets of trademark law were based on human consumers. The basic principles surrounding trademark law include; imperfect recollection, confusion, average human consumer, visual phonetic and conceptual similarities which are all characteristics of a human consumer. The registration of trademarks was important to avoid confusion of consumers. One of the requirements of registering a trademark is that the mark must be distinct. Confusion can either occur at the point of purchase or post-purchase. The law on trademarks tends to avoid confusion at the point of purchase by avoiding the registration of similar marks and the test is that of an average consumer. On the other hand, post-purchase confusion occurs after the goods have been purchased, and this is more likely to occur with the rise of online shopping. AI may recommend similar goods to what you searched for, and once the purchase is made, consumers may be dissatisfied with the goods due to their quality, or they may be counterfeit in some instances.
The law on trademarks envisages a situation where a purchaser goes to a shop and based on the distinctiveness of the mark, can purchase quality products. This then leads us to the question of whether AI can also confuse marks.
AI is programmed based on the information it is exposed to. Therefore, there is a lower likelihood of confusion as it depends on the algorithm and does not vary on a case-by-case basis. This poses a potential problem, especially since AI is still evolving and the predictive retail model is the future of purchasing consumer goods. The predictive retail model is characterized by the auto-replenishing AI (Curtis & Platts, 2017). This occurs where based on your shopping history, AI can replenish your supplies when they are about to be depleted. An example is the Amazon Dash Replenishment Service, which is programmed to replenish items such as groceries without human involvement. Further, the concept of the confusion of an average consumer formed a huge basis of the traditional law on trademarks. In determining the likelihood of a mark bringing confusion, the courts have adopted the test of an average consumer. However, with the rise of internet advertising and use of AI, the question that then arises is; who is the average consumer?
In the Google France versus Louis Vuitton Malletier SA (C-236/08) case, the court of appeal stated that the average internet consumer and the average consumer were the same, and the test applied was; they are reasonably well informed and reasonably observant and circumspect, the only difference is that one is an internet user. However, the question that still remains unanswered is whether AI meets the test of an average consumer. This is a particularly hard question that has to be asked for the law on trademarks to develop. Because with the 4IR there are more technological advancements that will develop and AI is going to be making decisions for humans and whether the same test will be applied, it will be very difficult to tell and may even render the law on trademarks useless.
In my opinion, as much as AI emulates the human mind, it is not human and can be too perfect sometimes. For instance, it has a powerful memory, where the chances of an AI confusing two brands is very unlikely, bringing us back to the point that AI will prefer one brand to the other not because of quality, but because of the information that it is fed.
Consequently, if AI is considered an average consumer, there are questions of biasness of AI that then arise because if the output given is dependent on the information that the developer feeds into the machine, then there are other brands that are likely to miss out on the recommendations given by the AI: unfair trade practices. Further, when the AI gives its recommendations, we are not certain that it can tell an original product from a counterfeit one. Also, it cannot tell whether a product is defective at the point of purchasing, leading to a lack of customer satisfaction. Therefore, determining whether AI can confuse various trademarks will be difficult, especially where infringement arises as a result of the use of AI to purchase goods.
Further, the trademark law is structured so that in the event of infringement, the person whose mark has been infringed against can bring a legal action against an infringer. However, with the rise of AI, where AI does the shopping for you or makes the recommendations, in the event of infringement, who is held liable? This question is related to the average consumer doctrine, as is the basis of determining infringement of trademark cases. This question arose in the case of Lush versus Amazon  where Amazon was held to have infringed on Lush’s trademark. Once a person searched Lush, Amazon would recommend other products other than Lush products. On this, the Court held that this amounted to infringement as Lush products were not being sold despite searching specifically for those products. This case brought out the bias that comes with AI, and in some cases, results in infringement. This is among the few instances where it was easy to find an infringer for the actions done by a machine, and may be a guide as to who is held liable in cases of AI infringement.
Therefore, in determining the nexus between AI and trademarks the following questions arise; is AI the new consumer? Can AI differentiate confusing marks? Can AI determine counterfeit products? Does AI give an equal opportunity for trade for all trademarks? Does the use of AI promote unfair trade practices? Is the advancement of AI leading to doing away of the law on trademarks? Can AI infringe on a trademark? Does AI portray biases in product recommendations?
It is not disputed that the 4IR is here with us but the unpreparedness that the policy makers, legislators, and courts have portrayed has led to these conversations. Artificial Intelligence makes our life easier, but we also have to regulate it. AI goes to the core of the law on trademarks as it is changing how products are being purchased. Therefore, when it comes to the law on trademarks, we need to recognize that the various technological advances affect the law on trademarks positively and negatively. Positively in the sense that AI programs such as ‘Project Zero’ is a notable good AI that is being used to identify counterfeit products and filter them out to avoid the continued infringement of trademarks. This reinforces the purpose of the law on trademarks.
On the other hand, AI recommendations of certain products are marred with biasness. Unlike in the past, goods recommended by AI are not being consumed because of their quality but because of how much discount they have or the shipping period or similarity with what the consumer purchases (Curtis & Platts, 2017). This then begs the question of whether AI is the new consumer, as it has completely taken over the purchase process in some cases. Consequently, the trademark law should be altered to factor in the existence of AI not just as an assistant but as the consumer because, with the emergence of auto-replenishing, AI will be making all decisions without human intervention.
In addition, bias will always exist in AI as the output depends on the information made available to the AI by the developer or programmer. Hence the legislators should develop policies against the unfair trade practices perpetrated with the use of AI and the effect it has on consumers. The use of AI does not encourage fair competition practices within the market, and in a way, this breaches the consumers’ rights, however, this is another lengthy discussion that cannot be addressed here.
Moreover, the traditional tenets of trademark law ought to be revised to factor in the emergence of AI. For instance, in the case of Cadilla Healthcare Ltd. v. Cadilla Pharmaceuticals Ltd (2001), the court stated that a consumer can only be considered an average consumer if they have average intelligence and tend to have an imperfect recollection. This is not the case for AI, because AI has a perfect memory as it has been programmed that way and does not possess average intelligence.
In conclusion, there is a need to recognize the effect of AI on the law on trademarks before the effect is so huge that it can no longer be controlled. Further, we need to have conversations around laws and policies that will recognize AI’s role in shaping trademark laws. The question still remains; Is AI the new consumer? The courts do not move suo moto, so the answer can only be found once they are faced with a case to this effect; for now we can only wait.
Cadilla Healthcare Ltd. v. Cadilla Pharmaceuticals Ltd (2001)
Cosmetic Warriors Ltd and another v amazon.co.uk Ltd and another  EWHC 181 (Ch)
Curtis, L., & Platts, R. (2017). AI is coming and it will change trademark law. Journal of Intellectual Property Law & Practice, 12(7), 532-537. <doi:10.1093/jiplp/jpx068>
Curtis, L., & Platts, R. (2020). Trademark Law Playing Catch-up with Artificial Intelligence? Journal of Intellectual Property Law & Practice, 15(5), 327-334. doi: <10.1093/jiplp/jpaa023>
Google France versus Louis Vuitton Malletier SA (C-236/08)
Gutierrez, S. (2022). Momentive study: Gen Z social media and shopping habits. Momentive, Inc. Retrieved from <https://momentive.ai/resources/reports/gen-z-social-media-and-shopping-habits>
Steponénaité, V. K. (2019). Alexa, are you confused? Unravelling the interplay between AI and (European) trademark law. Journal of Intellectual Property Law & Practice, 14(11), 812-821. doi: <10.1093/jiplp/jpz158>