Artificial intelligence (“AI”) is big news at the moment, so it’s no surprise that lawyers are considering the IP issues relating to AI. Some of the implications are really quite profound. In this article, I will discuss some of the copyright and trade mark implications. 

The copyright issues of AI have been discussed by a number of people. There has been a particular focus on the issue of who owns the copyright in AI-generated works, specifically works where there is no human input at all. As AI is not a legal person, this might present real problems. 

The article that has interested me most is one entitled “AI is coming and it will change trade mark” law by Lee Curtis and Rachel Platts, published by Managing IP. This article discusses the inherent conflicts between trade mark law and AI. To understand this conflict, we need to go back to basics.

To this day, a trade mark is described as a “badge of origin” or, sometimes, a “badge of control”. What this means is this: although a trade mark may not necessarily tell the consumer which company manufactured the product, it does tell the consumer that all goods featuring that trade mark were made by the same company, and will therefore be of a consistent quality. It allows the consumer to again buy a product that they have previously been satisfied with, and avoid buying a product that they have previously been dissatisfied with. 

So, trade marks and trade mark law are all about human interaction with brands. Just think of the concepts that form the basis of trade mark law: distinctiveness, reputation, confusing similarity (visual, phonetic, conceptual), imperfect recollection, and consumer propensity to concentrate of the first element of a trade mark rather than what follows.

This human interaction has always been there. Right from the days when shopkeepers guided us on purchasing decisions, to self-service shops, to online retailing and social media prompting. But AI is a game-changer because it’s predictive and it takes the human out of the purchasing process. 

So, just what does AI look like in this context? The authors tell us that we’re moving from an age of “shopping then shipping” to an age of “shipping then shopping”. In other words, one where AI orders goods for us, the goods arrive, and we then decide whether or not we want to keep them. Some examples of AI include:

• Amazon’s recommendations based on orders
• Amazon Echo. The authors point out that, although devices like this only recommend at this stage, it’s likely that they will move to automated buying in the not-too-distant future
• bots, for example, eBay’s ShopBot, which helps users find the best deals 
• Amazon Dash, which replenishes consumables and which could, in time, become completely automatic. We’re told that in due course, this could be connected to your fridge, so things like fresh milk and cheese will be ordered automatically 
• Mona, a virtual shopping assistant in the field of fashion, which can be set to alert users to styles and not simply brands 
• Pepper the robot, which reads moods in the retail environment and makes suggestions

So, retail will therefore stop being reactive, but many questions remain. Will AI consider branding at all or simply issues like price and style? Will the AI models allow for returns in the case of confusion? What place will there be in trade mark law for all those concepts of confusion (visual, aural and conceptual), blurring, dilution, the average consumer? Does AI become the average consumer? Does AI have imperfect recollection? Can AI be confused? 

There’s more – what if AI suggests counterfeits, look-a-likes or compatibles? What if, as the title of this article suggests, AI interprets the request to refer to a type of product rather than a particular brand name? Are there any issues of secondary infringement? The authors don’t know the answers, of course, these will be for the developers of AI, brand owners, lawmakers and the courts to decide. 

The authors do, however, have some thoughts on the issue of secondary infringement. Based on earlier cases dealing with issues relating to AdWords, they suggest that the developers of AI will only be liable for secondary infringement in cases where the brand owner has given notice of infringements. And, based on the very old UK case of CBS v Amstrad, where the manufacturers of tape recorders that allowed for tape-to-tape recording (thereby making life easier for infringers) were held to be not liable as contributory infringers because copyright infringement was not the main purpose of the equipment, they feel that the same principles will apply to devices created to make shopping easier rather than to facilitate trade mark infringement. 

The future of trade mark law looks very interesting indeed.