Artificial intelligence or "AI" is defined as the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. Computer scientists sometimes call AI "machine intelligence" to distinguish intelligence demonstrated by machines from the natural intelligence displayed by human beings. Leading AI textbooks define the field as the study of intelligent agents: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals. Colloquially, the term AI is often used to describe machines (or computers) able to mimic cognitive functions that human beings associate with the human mind, such as learning and problem solving. Although lacking a uniformly accepted and clear definition, AI generally involves four steps: collect data, run the data through an analytical model to predict, optimize the model and make decisions, then have the system adapt or learn.

As AI becomes commonplace, it is playing an increasingly large and important role in the American legal system. AI is a part of how attorneys practice law, is used by a diverse array of clients and even plays a role in the judicial process itself. From AI-driven document review to computer sentencing guidelines to automated decision systems, nearly every lawyer will encounter AI at some point in the near future (if they have not already done so). The legal profession is struggling to keep up with the technological advances inherent in AI, grappling with the fundamental elements of how AI works and trying to understand its benefits and drawbacks.

More specifically, AI is a tool that may provide attorneys, judges, and businesses with potentially new insights into trademarks. Like most computer programs, however, AI will for the foreseeable future require human knowledge and interaction to maximize the utility of AI applications to trademarks (as well as to other disciplines), and undue reliance on AI may result in unintended and unreliable consequences. Attorneys, judges, and businesses must understand the technology to analyze and apply the critical thinking needed to evaluate AI applications and other developing and evolving computer programs as tools. A September 2019 TED (technology, entertainment, design) talk titled "How humans and AI can work together to create better businesses" provides entertaining information about some of the experiences by businesses using AI, which may provide insights into what might be expected of AI in the trademark field. The TED talk is available here.

Others have already begun to develop insights into the impact of AI on trademark law. The U.S. Patent and Trademark Office (PTO) held a full-day conference on Jan. 31, 2019, titled "Artificial Intelligence: Intellectual Property Policy Considerations." (The PTO has also launched a page on its website, www.uspto.gov, that provides information on its AI initiatives, public notices and responses, AI-related events, and outside resources.) Acknowledging that AI is expected to produce a new wave of innovation and creativity while posing novel challenges and opportunities for intellectual property (IP) policy, the conference included six panels featuring IP specialists from around the world. The trademark-focused portion of the day was a panel presentation titled "Does AI dream of electric brands?" The trademark panel considered questions such as (1) how is AI being used to enforce trademark rights, (2) how will AI affect trademark protection and branding, (3) will AI change the likelihood of confusion and liability analyses, and (4) how will AI impact the branding of products and the protection of trademarks?

Trademark law considers marks in the context of goods and services. The keys to a purchasing decision for goods and services are the information available to the consumer when making their decision, and who helps the consumer to make their decision (e.g., a store clerk). Companies offer virtual assistants. AI may soon make purchasing decisions for consumers entirely "behind the scenes." Ironically, therefore, AI may give consumers less information. Some consumers are happy to delegate their purchasing decision to an AI agent even now; others may be more comfortable over time. (Although fear and uncertainty about AI may prevail now, at least among some people, acceptance and familiarity with AI, over time, will likely eliminate those concerns.) Many consumers want to delegate decisions sometimes, but not other times.

AI notices trends and prior decisions and makes recommendations or at least narrows options for consumers. As AI gets better at predicting, the consumer may be delivered a product the consumer ostensibly wants without even asking (a shift from suggesting through offering to providing!). Eventually, the consumer may interact only with an AI agent as to why the consumer does or does not want an item. Keith Weed, chief marketing officer at Unilever, was quoted in The Economist as saying, "We'll be having bots trying to influence your bots about buying our products." Rather than build an association between brand and product (by investing in advertising), it may be more helpful in the world of AI to build an association between functionality and brand by investing in back-room AI so that brands use technologies in the "right" way (e.g., Google and Amazon).

Conventional trademark law is all about human beings and human interaction, both with brands and in the purchasing process. But retail in the AI environment will be predictive rather than reactive. AI effectively reduces or, at its most extreme, completely removes the human being from the product suggestion and product purchasing process. Trademark law coped with the self-service revolution of the 1900s, the Internet revolution at the turn of the century, and, mostly, with the recent social media revolution. In their article titled, "AI is Coming and It Will Change Trademark Law," ManagingIP.com at 9-13 (2017), authors Lee Curtis and Rachel Platts ask: Can trademark law deal with the AI "fourth revolution," which is changing the retail process from "shopping-then-shipping" to "shipping-then-shopping"? In its extreme application, shipping-then-shopping completely takes over the purchasing decision and, given the likely increase in returns, will increase the need to assess trademark issues such as post-sale confusion. This AI revolution raises several interesting trademark law questions:

  • Can AI be confused (likely or actually)?
  • Does AI take the place of the average consumer or is AI a "sophisticated" consumer and does that assessment depend upon the type of product or service at issue? (It would seem that an AI's level of attention does not vary according to the product or service.)
  • Does AI emphasize the beginning syllable of a trademark rather than the end, or the logo over the words in a composite mark, for a likelihood of confusion analysis?
  • Does AI have any bias in analyzing marks?
  • Will AI consider brands at all - or will AI simply focus on price, taste, nutritional information, availability, style, speed of delivery, quality, and the like when making a purchasing decision?

Consider the applicability to an AI agent of secondary infringement liability under conventional trademark law, i.e., is the agent or entity behind the agent liable for a consumer's confusion or poor decision? More specifically, when your Amazon Echo suggests and buys a product for you that infringes a registered trademark or is a counterfeit, does Amazon become a secondary infringer? Perhaps not, but the entity may be required to have takedown procedures.

Counterfeits are an increasing problem, with counterfeit goods comprising an estimated $461 billion or 2.5 percent of all global trade. AI offers hope in addressing the problem. For example, AI can help to identify counterfeit products and remove them from purchasing options. Amazon has a brand registry program to help brand owners. How can AI be further leveraged to solve the age-old problem of counterfeiting?

Given the future of AI, perhaps brands will become much less relevant to the purchasing decision. See J. Herrman, "All Your Favorite Brands, From BSTOEM to ZGGCD/How Amazon is causing us to drown in trademarks," The N.Y. Times (Feb. 11, 2020) (as used by Amazon sellers with minimal conventional marketing, pseudo-brands are challenging what it means to be a brand and, for many categories of products, sellers simply use Amazon's brand name and platform to sell their products). Regardless, it is quite possible that many of the long-held principles of existing trademark law will become irrelevant in the age of AI or will at least need to be applied in a new way.

A report from Stanford University and New York University researchers commissioned by the Administrative Conference of the United States and issued on Feb. 18, 2020, summarized the PTO's AI activities. The PTO has been experimenting with using AI when examining applications to register marks, and although such automation offers "substantial" benefits, the report identifies risks about due process and labor. Among the benefits to the PTO are handling searches and classifications so that trademark examiners can focus more on analysis. The PTO has "experimented" with AI and machine learning tools to automate classification, and with prototyped models for searching for prior marks, according to the report. These changes may "significantly improve the trademarking process" in the future. So far, however, the implementation has been "suboptimal," with the report citing issues like duplicate images and text identification.

On a broader scale, the report cited three main concerns for AI usage at the PTO. First, AI raises due process concerns, particularly concerns that search results would be harder to decipher, which would violate requirements for decisions to be explainable. Second, the unions that represent trademark examiners could be resistant to AI if there is any reduction in hours or other employment effects. Finally, the report warns that applicants may try to game the system by drafting their applications in ways that they know will avoid prior marks. The report suggested creating a rule to clarify the duties and obligations of applicants to minimize this risk.

There is great consensus and little doubt that AI will revolutionize trademark prosecution and enforcement over the next few years. According to research summarized in the Hogan Lovells Brand Benchmarking 2018 report, which targeted over 200 brand owners from a wide range of industries, a resounding 93% of businesses feel positive about the new technology. Time and cost savings are expected to be the biggest benefits of AI, which will impact trademark prosecution clearance searches, according to 93% of respondents. With respect to enforcement, AI is expected to facilitate online infringement searches and the preparation of take down notices. Although there is some concern about job security, the majority of businesses surveyed do not see AI as a threat. Commenting on the findings, Lloyd Parker, Asia Pacific and Middle East Head of Intellectual Property at Hogan Lovells, said: "There is a great opportunity for brand owners to use AI to gain efficiencies, speed up their work and streamline processes, while reducing costs and ensuring resources are used effectively. However, there is [a] worrying lack of awareness about AI, and businesses risk missing out on its benefits due to insufficient knowledge and investment in the new technology. This is an area where all companies should be paying more attention and seeking out beneficial opportunities."

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.