It is the year 2050, and the last trademark litigator has retired

It is the year 2050 and the last trademark litigation attorney has just retired. For several years, he had remained a lone outpost of human intervention in a field that had been taken over by Artificial Intelligence (AI) and advances in neurological science. An enterprising journalist, eager to preserve the story of how technology had changed the face of trademark litigation, sat with our recently retired attorney and asked him to recount the highlights. The journalist has shared an advance copy of the story with IPKat (now in its 47th year of publication). Here is the story, in brief.

“By the year 2030, the threshold question that engaged the legal system was whether it should continue with the time-honored system of relying on a human judge sifting through the evidence, listening to testimony, and pondering over written submissions, or should it focus on developing technological means to replace this process. The lack of certainty inherent in judge-made determinations of likelihood of confusion was well-recognized. Human courts had worked up a number of legal tests, each composed of a number of circumstantial criteria, which, taken together, were meant to guide the judge in reaching a decision.

But how this worked in practice, well, that was a matter of the quirks of the particular judge. All of this made for great theatre, what with the suspense of uncertainty, but whether this was what justice was supposed to do was quite another story. Maybe neuroscience, together with AI and the crunching of huge quantities of data, could come to the rescue.

As for neuroscience, the major breakthrough took place in 2035, when researchers identified the part of the brain that controlled cognitive processing of both two and three-dimensional signs on goods and services. Refinements were made in the research over the remainder of the decade, especially in the ability to go beyond laboratory settings and fashion the means for evaluating subjects in real-life settings, either in the physical purchasing environment or on-line. As a result, by the year 2040, it was possible with 99% percent accuracy to determine whether the subject was confused as to source of a branded product, having regard to a pre-existing brand.

These results seemingly guaranteed much greater certainty than did conventional judicial methods to determine whether there was actual confusion. But actual confusion constituted only a sub-set of the legal potential for establishing trademark infringement. Wasn’t the Promised Land for trademark litigators the much broader category of likelihood of confusion? Here, neuroscience did not seem to be able to offer much assistance. So researchers looked to AI and the development of algorithms that could ensure more valid and reliable results. WIPO duly put out a tender for the funding of up to four competing groups, each charged with developing an algorithm that was meant to yield better results in determining whether there was a likelihood of confusion, given a certain set of facts and applying the appropriate legal considerations. The jurisdictions to be considered were China, the US, the EUTM, India, Germany, France, England, Mexico and Brazil.

The question of how to determine the efficacy of the algorithm was front and center. The proposed approach was to base the algorithm to all decisions given since the year 1950, in which the finding of the trial court on likelihood of confusion was reversed on appeal. The winning algorithm succeeded in reaching a valid and reliable result in 97% percent of the cases that in practice had been reversed on appeal. Here too, refinements continued apace, until, by the year 2045, no more appeals from a decision on likelihood of confusion were filed in any of the relevant jurisdictions. The algorithm was, more or less, always reaching the correct result at first instance.

So what was left for the lawyers to do? In the main, in responding to a court-fashioned questionnaire, they simply submitted the requested information. At the beginning, there were disputes over whether these submissions were proper and, for a while, the successful trademark litigation attorneys were those who best gamed the system in responding to the questionnaire. However, once again, with refinements to the algorithm, the disputes diminished and ultimately, in effect, disappeared.

To be a trademark litigator meant being the gatherer of the information and documents requested. Arguments, whether over points of law, or mixed questions of law and fact, were taken care of as part of the questionnaire and the application of the algorithm. One observer noted that if, in prehistoric times, mankind had progressed by being a food gatherer, by the year 2050, gathering meant data and information to enable the algorithm to do its job. Ultimately, such gathering of information and documents became a mere utility; there was nothing left to be gamed and, with it, the job of a trademark litigator ceased to be relevant.”

The full story of the extinction of the trademark litigator function will appear in the July 4, 2050 issue of the journal. To read it, just give an oral request for the journal issue and it will be delivered to your personal device.

By Neil Wilkof

Picture on top right by Avtech Scientific and is licensed under GNU Free Documentation License.

Picture on lower left by Heinrich Harder (1858-1935) and is in the public domain.