Can artificial intelligence transform trademark law?

The main reason I write the daily observations is because I want to know where I’m wrong. — Ray Dalio

This is the first in a series of posts where I am trying to follow Ray Dalio’s example and write about where I think algorithms can be of most use to the legal system in the next 2 to 5 years. Many of the recent articles on how AI is going to “transform” the law have played it safe and focused their predictions on how things will look many decades from now. I want to talk about how better algorithms can help now, or at least, soon.

This is, inevitably, a work in progress as I try to develop the right questions to ask. There does not seem to be any doubt that a whole collection of technologies under the heading of “Artificial Intelligence and Machine Learning” are going to transform EVERY aspect of life, and it seems to me literally impossible that the legal system will be exempt from the coming changes. But to date, lots of people writing about “the coming transformation” seem to want to jump past the “how are we going to get there?” part.

In any case, I’m going to start — somewhat arbitrarily — with trademark law. And, I propose that, when we try to think about how algorithms might change the practice of trademark law, some basic “opportunity assessment” type questions. From this I lean on the approach developed by the folks at Pragmatic Marketing

Step 1, what are the problems with trademark that the CURRENT legal system solves?

Step 2: how pervasive are these problems? How urgent are they? Are people willing to pay to solve them?

Step 3: how do people solve these problems now? And what are the benefits and costs of a different approach, specifically an approach that incorporates machine learning.

What are the problems that a trademark lawyer helps a client solve

My initial hypothesis — and here I’m hoping my trademark lawyer readers are going to jump in and correct what I got wrong — is that a lot of the heavy lifting in trademark law comes down to, essentially (and as we’ll see below I’m oversimplifying) figuring out  whether one trademark is so similar to another that it INFRINGES on the other mark, i.e consumers will be confused. And a lot of what a good trademark attorney does for you, as a business owner, comes down to ASSESSING whether a COURT will rule that one trademark is close enough to another that is is said to “infringe” on the other mark.

Say, for example, that a small coffee roaster in New Hampshire sells a dark roast called “charbucks”. Does that name remind you of any other brand you might have heard of? Is that name confusingly similar to any other coffee brand you might have heard of?  Not surprisingly Starbucks thought so, and they sued. But ultimately the court held that there was no infringement.

This case is a good illustration of a few points about trademark law. First, the issue is not, is the name similar to another name, or at least that’s not the whole issue. The legal question is, would your average consumer be confused. Would the consumer think “hmmm hey how about that ‘charbucks’ well, I bet that is made by Starbucks, and I like Starbucks so I’m going to buy charbucks?”

Trademark basics and understanding the moron in a hurry

Quick reminder from trademark 101 (which I got a B in, so you should only be about 85% confident that I have this right) a trademark is NOT just a name, but a name (or a logo or other mark) associated with a particular BUSINESS, well, technically with a particular good or service or set of goods and services. The somewhat shopworn illustration of this principle is that both Delta Airlines and Delta Faucets use the “Delta” mark, but there is no confusion because one sells faucets and the other operates an airline. (By the way, I looked for, but could not find any examples of Delta Airlines and Delta Faucets sending angry letters to each other. Note that Delta airlines wound up with delta.com domain name. I still wonder if they sometimes get each other’s mail by mistake….)

But what about the case of Apple. You’ve heard of Apple, right? It was founded in 1968. By a band called the Beatles. Then, in 1976, somebody in California started another company, called Apple Computer. The Beatles were not pleased, and they sued. The parties reached a settlement because, after all, computers and music were totally different businesses,  so there was no chance for confusion.

So what counts as confusion? Well, of course that’s the $64,000 question. The court cases are themselves confusing. My favorite version of the “what counts as confusion” test is Lord Denning’s “moron in a hurry” test https://en.wikipedia.org/wiki/A_moron_in_a_hurry.  After a British newspaper launched a new paper called the Daily Star, the British Communist Party sued, claiming that the name Daily Star was confusingly similar to their paper the “Morning Star.” Lord Denning threw the case out, explaining that, if you put the two papers side by side, “only a moron in a hurry” would be confused.

The Courts in the U.S. never adopted the “moron in a hurry” test (more’s the pity) but have instead created several different versions of a “multi factor test” for determining infringement. And in fact, federal courts in different regions have adopted slightly (more maybe more than slightly) different versions of the tests to apply to determine whether there is infringement (when federal courts in different regions apply different standards, this is referred to as a circuit split).

In a very helpful paper, An Empirical Study of the Multifactor Tests for Trademark Infringement, Barton Beebe studied several hundred trademark decisions and concluded, basically, that the courts were kind of all over the map.

So, where does that leave business owners seeking help with trademark law issues? First, again this is a hypothesis, there are two market segments. First there are owners of EXISTING marks, i.e. Starbucks or Mattel. And there are businesses who are CONSIDERING adding NEW marks. The hypothesis I want to explore in the next post is about two problems: (1) how to tell, from millions of trademarks out there, which ones are relevant, i.e. POTENTIALLY conflicting with your trademark and (2) for a given trademark, how to assess the strength of your case against the other party.

More on this next time.

Advertisements