The subject of artificial intelligence is presently very much a hot topic. It has the potential to impact on all aspects of the way we live and the economy leading to significant changes in how we live, how we work and importantly and possibly often overlooked at times, how we search for and buy products.
Artificial intelligence has the potential to have a major impact on the way fashion products, in particular, are searched and purchased in the future. Although one might presume the way one purchases a dress, a scarf or a pair of trousers have been constant over time, that is not case. The way consumers search and buy product has changed over the decades, and artificial intelligence and its applications will further impact on the purchasing process.
One only has to look back in history to see the way fashion products have been bought has changed. In the nineteenth century, the human shop assistant was key and an important filter between the consumer and the product and brand. The human shop assistant provided information on what product was available and recommended what to buy. We then had the rise of supermarket and the department store, where ‘self-service’ became a more important aspect of the way fashion products were purchased. There was no filter between the consumer, the product and the brand. The consumer made the purchasing decision alone and was aware of all the product in the store, as it was on display in front of them. With the rise of the internet, the way consumers bought fashion products changed again, with the amount of information available to the consumer increasingly expeditiously and then we had the rise of social media, when other influences started to impact the way consumers bought product, such as the ‘likes’ of friends and family and the rise of influencers in the form of celebrities. We now have the rise of AI.
Artificial intelligence applications have the potential to provide a new filter between the consumer, the fashion product and the fashion brand. Artificial intelligence applications include applications such as product recommendation applications such as those which have been available for years on Amazon, customer service chatbots, personal assistant applications such as Alexa, Siri and Cortana, image search assistants such as Amazon Echo Look, shopping product search assistants such as eBay’s Shopbot, replenishment applications such as Amazon Dash and robotic customer assistants such as Pepper.
If one considers the basis of trade mark law it is all about the interaction between the consumer and the brand, and importantly what information is available to the consumer in the marketplace and how and importantly who makes the purchasing decision. Further, trade mark law is fundamentally based on human frailty. If one thinks of the ‘buzz words’ of trade mark law, such as confusion and association, the phonetic, conceptual and visual comparison of brands, imperfect recollection and the slurring of trade marks these all consist or are related to issues of human frailty and limitations. However, when you take the ‘human frailty’ from trade mark law with the introduction of AI applications what is the result?
Can an AI application such as Alexa be confused? Does Alexa have imperfect recollection? Does Alexa slur trade marks? Who is the average consumer when AI applications are involved in the purchasing decision or indeed making the purchasing decision. Although not widely used at present, AI applications such as Alexa have the potential to purchase product on behalf of consumers with little or no consumer interaction in the buying process, often based on past purchasing decisions and other variables such as price, quality and colour. What impact will this have in trade mark infringement cases, where the human consumer may not be doing the purchasing? However, at least in the short to medium term, the impact of AI applications will be greater, at least when one considers its impact on trade mark law, when one considers the question of what information is available to consumer when an AI application is used. By definition the AI application is aware of the product information when considering the purchase, not the consumer. An AI application such as Alexa on average makes three purchase recommendations to the consumer, but the consumer is not aware all the products that were available to the consumer. This has major implications for trade mark law, as fundamentally based on the interaction between the consumer and the brand and how and importantly what product and brand information is available to the consumer.
AI applications also raise important questions of liability. If an AI application purchases an infringing product who is liable? Although there have been no cases to date explicitly on the question of the liability of AI applications in intellectual property disputes, some European Union case law does provide pointers on this issue. In the Louis Vuitton vs Google France case which concerned the automatic processes associated with AdWords generation in the Google keyword advertising function, the Court of Justice of the European Union (‘CJEU’) found that Google France was not liable for trade mark infringement unless they were found to take an active role in the keyword advertising generation and were aware of infringing activity.
Further, in the case of L’Oréal vs eBay which concerned the liability of online marketplaces in the sale of counterfeit products, eBay was found not to be liable unless they were aware of infringing activity and did nothing. To a degree these findings have been re-confirmed in a recent opinion of the Advocate general in the Coty vs Amazon case. The issue of liability therefore seems to be heavily influenced by knowledge of infringing activity. It would therefore seem that the owners of AI applications such as Amazon would not be liable for any infringing activity that an application such as Alexa might stumble into providing they were not aware of the activity pre-purchase and had in place sufficient take down procedures, similarly to those applied to keyword advertising and online marketplace platforms. Further, a platform such as Amazon also has the means to control the market ‘bio-system’ in which the AI application operates and does employ their own anti-counterfeiting AI applications to reduce any potential for liability.
Also the issue of copyright and design law is of importance to the fashion market and AI has its own impact in this sphere. AI applications such as Amazon Echo Look can search for fashion items by look and further eBay’s Shopbot application can search the eBay platform looking for products on sale with certain design features such as a specific colour, similar features are also available on social media platforms such as Pinterest. Would the provider of an AI application such as eBay’s Shopbot be liable for design or copyright infringement if it were to refer a customer to counterfeit product by mistake? If the reasoning detailed above is to be followed then the answer again would be no providing the AI application provider was not aware of the infringing activity and had sufficient takedown procedures in place.
The advance of artificial intelligence in the world of fashion commerce and retail potentially opens new areas of business for consumers and fashion manufacturers and retailers, but also potentially new areas and issues for the law to deal with. The future will be no doubt be interesting from both a fashion and legal perspective.
Lee Curtis is a Partner and Chartered Trade Mark Attorney at the Manchester Office of HGF in the United Kingdom: http://www.hgf.com/about-us/our-people/lee-curtis/