Guest Article: AI is Changing How Products are Bought and by Definition Will Change Trade
A guest article written by Lee Curtis and Rachel Platts, HGF
At its very basic level, trade mark law fundamentally is about how the public buy products and their interaction with brands. After all, a trade mark is defined as a sign which distinguishes the products of one undertaking from another undertaking, but where does this ‘distinguishing’ happen?
The purchasing decision
The ‘distinguishing’ happens when the public consider products carrying brands pre-purchase, purchase products carrying those brands or when they evaluate products carrying those brands post-purchase. As a result, any form of new technology which impacts that purchasing decision also impacts trade mark law. Artificial intelligence (‘AI’) is, and will have, a large impact on how the public buy branded products and thus trade mark law will have to adapt to reflect those changes. Trade mark law is ultimately the most practical and, historically, one of the most adaptable forms of the law. Indeed, over time trade mark law has been remarkably adaptable to the changing commercial landscape. However, how will AI impact the purchasing process? In the Harvard Business Review(May 2019), Nicolaj Siggelkow and Christian Terwiesch postulated that advances in technology, most notably new forms of AI, means that brand owners are no longer waiting for customers to come to them to buy their products. A brand owner will no longer simply place products on the market and wait fora consumer to see their brand and buy their product. AI enables a brand owner to predict the needs and desires of the consumer. Siggelkow and Terwiesch identified four new strategies where AI has the potential to affect significantly and these strategies are the so-called ‘respond to desire’, ‘curated offering’, ‘coach behaviour’ and ‘automatic execution’ strategies of product offering.
‘Respond to desire’
AI enables brand owners to respond to desire. Probably the best example of the respond to desire strategy on the market at the moment is the Amazon Dash service and other automated replenishment services, such as wifi enabled applications in Brita water filter machines, which automatically orders product when the consumer uses the branded product up. In time, such applications will be linked to the ‘internet of things’ so for example your fridge will predict and subsequently order your food-product desires and you will never run out of household essentials such as toilet roll.
The curated offering strategy has been in place for years in a form AI that is often overlooked, notably the product recommendations which retailers such as Amazon provide when buying books and other products. This strategy has also been applied by Netflix via their film and programme recommendations based on past viewing history and food delivery services such as Blue Apron in the US and Hello Fresh and Mindful Chef in the UK.
The coached behaviour strategy is a variant of the curated offering and respond to desire strategies, but more involved in customer interaction and perfect for a customer who in the words of Siggelkow and Terwiesch needs a ‘nudge’ or ‘suggestion’. AI will be able to monitor the needs and desires of the consumer and then suggest products and potentially brands to the consumers to purchase in response to that information, which a consumer may or may not act upon. In many ways, this is how Amazon’s Alexa, in a default form, works at present in product purchasing. Alexa makes suggestions for products in response to a query from the consumer, and those suggestions are based on the consumer’s past purchases. The consumer can then decide whether to purchase or not based on those suggestions. Companies such as Thread have applied this strategy to fashion purchasing in the UK. However, the consumer is effectively ‘coached’ in this scenario, as they are not completely aware of all the products on offer, which Alexa or Thread is searching, nor are they completely aware (or remember) their past purchases which impact on the product suggestions made by such AI platforms.
Then we come to the most radical strategy: ‘automatic execution’. Automatic execution has the potential to turn the purchasing process on its head. In a further article in the Harvard Business Review (October 2017) authors Ajay Agrawal, Joshua Gans and Avi Goldfarb likened automatic execution as turning the traditional purchasing process of a ‘shopping-then-shipping’ model to a ‘shipping-then-shopping’ model. When AI is sufficiently adept, it will predict, order and deliver products to the consumer with no interaction from the consumer in the actual purchasing process whatsoever. For this retail model to work, the AI must get to a sufficiently accurate level of prediction to avoid a large level of returns of product, an issue already bedevilling online fast fashion retailers such as ASOS in the present. Further, it is likely that a reasonably large proportion of the purchasing public will never have sufficient trust in AI to effectively sub-contract all, or a proportion of, their purchasing decisions to an AI Assistant, but many will as the trust in AI rises.
What does this mean for trade mark law and practice?
When a consumer buys a branded product fundamentally there are two key processes involved which impact on trade mark law. Firstly the assessment of brand information on the market by the consumer pre-purchase and then secondly the purchasing decision itself.AI impacts both these processes to a greater or lesser degree. With an ‘automatic execution’ or ‘pure AI’ model of retailing product, the consumer is completely removed from both stages of the purchasing process. If you think of many of the concepts of trade mark law they are based on human frailty, for example, imperfect recollection, consumer confusion, the slurring of trade marks by consumers, indeed what of the conceptual, visual and phonetic comparison of trademarks? Will an AI assistant have a perfect memory? Will an AI assistant have perfect pronunciation? Will brand owners have to provide perfect pronunciations for AI voice assistants? Will an AI Assistant ever be in a rush? Will the fact that one product is worth pennies and others thousands of pounds have any impact on an AI assistant in making an assessment of trade marks in a purchasing decision? Indeed can an AI assistant ever be a ‘moron in a hurry’ in the words of one English IP judge? Who is the average consumer when AI is involved in the purchasing process, indeed is the concept of average consumer simply redundant? Indeed, with some the purchasing decisions which an AI assistant engages not involving brand considerations at all, but the colour, price and quality considerations based on past purchasing decisions of the consumer. In these situations is there trade mark infringement at all, even if the AI assistant buys infringing product? This may also impact in the consideration of the level of damages in trade mark infringement proceedings, which at least in the UK are linked to the benefit or compensation for the use of the infringing trade mark. What happens when the purchasing decision made by AI is not driven by brand considerations at all? Further, what of liability in trade mark infringement proceedings when AI is involved? We have raised many questions above and not provided many answers. Here however, we feel case law with regards liability in keyword advertising and online marketplaces can provide some answers. In the Google France keyword advertising cases and in the L’Oréal v eBay International online marketplace case, Google and eBay were not held liable for infringing activity which they did not take an active part in. Unless a defendant is on notice of infringing activity it is likely that similar reasoning would be applied to an AI assistant which had bought an infringing product. However, the owner of an AI assistant, i.e. the company behind the assistant such as Amazon with Alexa, will be held liable if they have active knowledge of the infringing activity. Indeed, Professor Luciano Floridi (a professor of philosophy and ethics of information at the University of Oxford) has suggested that one might look to Roman slave law for guidance on the issue of liability, where the concept of the ‘intelligent slave’ under Roman Law transferred liability to the owner. Applying this to AI, the owner/designer of the AI system would be the owner of the intelligent slave, and thus liable. Further, AI does raise the further possibility that there will be a record of how a product is purchased and why and how an AI algorithm works could potentially become evidence in trade mark infringement cases. The increasing use of AI does herald change in the way products are purchased. However, as we outlined at the start of this article trade mark law historically has been highly adaptable, so there is reason to believe it will simply adapt again in light of the AI revolution. Indeed, the way product has been purchased over the last century has not been static. If one looks to a typical nineteenth century store, it would have involved a human shop assistant with the product information available to the consumer being relatively limited, almost akin to the ‘coach behaviour’ retail strategy outlined by Siggelkow and Terwiesch. When the modern supermarket was invented in 1916 at the Piggly Wiggly store in Memphis, Tennessee we moved away from that coached behaviour to a structure where the human consumer made all the decisions and was aware of all the products on offer in the store. The importance of the visual comparison of trademarks as a result increased. The internet and social media revolutions heralded new forms of how product was made available to the consumer. Ultimately, trade mark law has adapted to these changes and we will simply have to wait and see exactly how they will impact trade mark law in weird and wonderful ways.
About the authors: Lee Curtis is a Partner and Chartered Trade Mark Attorney at the Manchester Office of HGF.
Rachel Platts is a Trainee Trade Mark Attorney at the Manchester Office of HGF.
Guest Article: Artificial Intelligence and IP - Mapping Legal Challenges for the European Digital Single Market
A Guest Article written by Giancarlo Frosio, Senior Lecturer and Researcher at the Center for International Intellectual Property Studies (CEIPI) at the University of Strasbourg.
The regulation of Artificial Intelligence (AI)’s activities is set to become a primary policy issue. Virtual agents, sapient algorithms, robots, will have a terrific impact on the European Digital Single Market (DSM). Artificial Intelligence (AI) and robots have been the subject of science fiction for some time. That fictional future is now a present reality. The AI market is predicted to grow from $8 billion in 2016 to more than $47 billion in 2020. Investment in AI increased more than 300 percent in 2017 compared to 2016. Intelligent machines, machine learning algorithms, sapient bots and neural networks have invaded our daily life. The digital society will be increasingly characterized by the interaction of human actors and non-human technological actants or virtual agents within the so called "infosphere". In this context, there is a need for a policy framework that can promote a balanced coexistence of actors and actants in the DSM, so that EU citizens may reap the benefits of disruptive technologies and innovation rather than being overpowered by them. In particular, the so-called Forth Revolution is also a revolution where machines come as innovators and creators. At least five themes are relevant for legal practice and research in this domain: IP protection for AI technology, regulation of information and data used as inputs for AI, ownership and protectability of AI’s output, Digital Right Management (DRM) and IP enforcement through AI.