AI is posing unique challenges to tech companies. Claire Hutson from Marks & Clerk shares why a strong portfolio of trade marks is essential for any firms involved with AI. 

Whether it fills you with fear, excitement or both – there's no denying that artificial intelligence (AI) is a growing trend. 

AI start-ups are popping up all the time with some becoming multi-million dollar companies in just a few years. 

For new companies, having a strong portfolio of AI specific trade marks is an asset that can help attract and persuade investors. 

For established companies, making sure that they have a strong base from which to enforce their trade mark rights is essential, especially if that trade mark frequently makes global headlines.

The UK is frequently cited as the European hub for AI; however, France, Germany, Spain and Sweden all have active AI industries. We have seen that more and more trade mark applications being filed in the field of AI. 

In the five years to February 2019, nearly 45,000 trade marks were filed at the EU Intellectual Property Office covering AI. 

17,000 of these came from the UK, suggesting the UK is one of the largest centres for AI software companies at the moment. 

There are also a growing number of trade mark applications covering AI in Scandinavia, with 3,500 applications filed in Norway, Sweden, Denmark and Finland combined.

For those companies that have filed trade mark applications for their marks covering AI software there are a number of unique challenges to consider.

Drafting specifications

Drafting specifications must be considered very carefully. 

'Software' as a term alone is currently accepted in Europe however, it is no longer practically sufficient to cover this huge area. 

Including specific detail about the nature of the software is now standard practice and this is especially true for AI. However, only including terms that specify AI may not be enough anymore either.

So, what terms should be included?

Being overly technical can create problems at trade mark offices. While terms such as machine learning, deep learning and tree search may be familiar to programmers, they may be less so to trade mark examiners.

We have found examiners in territories ranging from the EU to Japan have objected to these terms. Is it sufficient that a person "skilled in the art" (to borrow language from our patent attorney friends) would be familiar with these terms when a business not active in AI may not? So far, we have found that it is but this may not continue to be the case and is not standard across territories.

Alternatively, you may choose to specify by the industry the software is to be used in. 

AI has applicability in any number of industries including healthcare, fintech, e-commerce, energy, autonomous vehicles, cyber security and more. 

As such, it can be difficult for a company to know early on if they might enter one of these areas in the future. It is also difficult to predict industries that may utilise AI in the future; some would argue that no industry will be free of AI in a few years.

It is essential therefore that trade mark attorneys drafting specifications for AI companies understand the nature of the business and how that might change in the future.

Is the trade mark in use?

A second issue that can arise is a lack of evidence.

Many AI companies are active in researching software and develop new algorithms to solve specific problems. 

It is also very common for large tech companies to acquire AI start-ups and incorporate the AI developed in-house into their products. 

The resulting AI software may be used in many consumer products e.g. Amazon's Alexa or Apple's Siri. The problem arises in that these products are made up of many parts, that do include the background AI software, but the consumer is generally not aware of this. 

Many of these companies become very well-known as they are frequently the subject of news articles. However, when enforcing their trade marks these companies may need to rely on evidence either to defend a cancellation action, oppose or invalidate an application or to prove that it is being used to allow an application to be registered in the US.

How does such a company then show that the mark has been used in relation to AI? 

Without standard forms of evidence such as invoices or advertising materials that clearly show the trade mark, this can be tricky. 

If there is enough linking evidence to show that a product has incorporated AI this may be enough but this is not always available. The question then becomes: are news articles and social media content enough to show use of a trade mark?

In our experience, some trade mark offices are more accepting of unusual evidence. The Chinese trade mark office, for instance, has previously accepted large amounts of press articles to support oppositions against applications for similar marks for both similar and dissimilar goods and services. 

Likewise, Campbell Soup's recent filing of the term 'Chunky' in the US relied heavily on media mentions and pop culture references.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.