The increasing use of artificial intelligence (“AI”) has raised many legal and ethical concerns including in the area of trademark law and practice. While there are benefits to the use of AI, the technology also presents new intellectual property challenges, particularly in connection with U.S. trademark application policies and procedures.
AI has the potential to transform the trademark application process in many ways: AI tools can enhance clearance search capabilities, identify potentially conflicting marks, and streamline the application process. However, these technological advancements also raise several concerns about accuracy and whether they can capture the complexities and nuance inherent to the practice of trademark law.
Potential for Inaccurate Trademark Searches
A key step in the trademark application process is conducting a clearance search to determine if the mark is already in use and to assess any infringement risks. The United States Patent and Trademark Office (“USPTO”) offers applicants and practitioners the ability to search the federal trademark database for similar marks before deciding to move forward with an application. AI tools used in the search process can present challenges to providing reliable and accurate search results.
It is reasonable to want to use AI tools in a clearance search. Such tools can scan large datasets and, in theory, quickly identify conflicting trademarks. However, AI algorithms have not yet fully captured the nuance of trademark analysis. For example, the algorithms may not comprehend the trademark “use in commerce” requirement, namely whether the trademark is used in the marketplace and in connection with the claimed goods and/or services. A “use in commerce” analysis is not a precise science; it considers several factors including the nature of use, promotional/marketing efforts, sales, and marketplace impact. Determining whether “use in commerce” is sufficient and not merely token use can be a complex analysis that AI tools might not have the ability to perform.
AI-driven search results may also lack relevant cultural context and skew trademark analysis. The algorithms may overlook cultural sensitivities that inform the context under which a certain phrase, slogan, or image is deemed inappropriate. An AI model may find a trademark featuring cultural elements acceptable, whereas human review would investigate further as to whether the trademark is offensive or exploitative. While AI is arguably better suited for identifying similarities in logos and word marks, it has not yet learned the role of cultural sensitivity and competence. This in turn makes it less dependable for comprehensive trademark analysis.
Risk of Misrepresentation
USPTO application filing procedures require applicants to provide various documentation in support of a trademark application. Such documentation includes a drawing of the trademark, verified statements, and specimens (evidence showing use of a trademark in commerce), many of which are submitted under penalty of perjury which carries legal consequences for making willful false statements.
While AI’s ability to streamline this process could make it easier and more cost-efficient to prepare and submit trademark applications, applicants may be submitting applications without meeting the requirements and understanding the consequences of failing to do so. For example, AI-driven tools could lead applicants to prepare inaccurate goods or services identifications. USPTO rules and guidelines require applicants to specify the goods or services they intend to, or currently offer under the applied-for trademark. An AI tool could select the wrong identification and class of goods/services and not adhere to the standard wording from the Trademark ID Manual. This misrepresentation before the USPTO, could lead to a rejected application and with recent changes in the trademark fee schedule taking effect January 18, the USPTO will now charge a $200 per class fee for applicants who do not use the standard wording. The recent fee changes also bring a $100 per class fee for insufficient information in a trademark application, information that AI tools might not account for in the automation process.
Moreover, AI has had an increasing role in the creation and submission of specimens. AI systems can create fake specimens including website mock-ups, product labels, and product packaging that appear legitimate. A primary risk associated with using AI-generated specimens is the potential for fraud or misrepresentation before the USPTO, which carries legal consequences for applicants. While Examining Attorneys have become more adept at identifying fake specimens, as AI tools become more advanced there is a greater potential for misleading the Examining Attorney into believing there is legitimate and sufficient use in commerce when there is not. As a result, trademark applications may be granted registration without meeting the necessary requirements, which in turn compromises the integrity of the examination process and limits ownership opportunities for legitimate applicants.
Ownership
AI tools have also been used to create logos and slogans, which for many brands can be another cost-saving measure. However, the AI-generated logos and slogans could contain protected elements from other trademarks or outright duplicate them. This not only creates an infringement risk but raises questions about ownership and eligibility for trademark protection.
Additionally, many AI-generative tools have terms and conditions controlling use of the AI output. When users sign up to use an AI tool, they are often required to accept the platform’s terms and conditions which may contain language granting the provider certain rights over the AI-generated output. For example, some platforms may claim ownership rights to the output and set use restrictions including preventing users from using the output for commercial purposes. Other terms and conditions may grant the platform an exclusive right to license the output. Such conditions undermine a user’s ability to build, maintain, and assert exclusive rights in the trademark or logo. For businesses this could mean limited IP rights, the inability to commercially exploit its IP, and expensive ownership disputes.
Ethical Obligations
In April 2024, the USPTO issued guidance on the use of AI in practice before the USPTO. The guidance clarifies what constitutes acceptable AI use in preparing and prosecuting trademark applications. Among the key points, the USPTO emphasized that practitioners remain responsible for the accuracy of all submissions regardless of whether AI tools were used. Generating and submitting fake information in an application or fake specimens that do not legitimately represent use of the mark constitutes unethical conduct and could result in legal consequences including sanctions, suspension from practice, and even criminal liability.
Notably, this USPTO guidance acknowledges that AI can be a valuable tool for efficiency and has the potential to lower the barriers to practice and the related costs. However, the value of these AI tools does not outweigh a practitioner’s responsibility to ensure accuracy in the filings and the responsibility to maintain integrity in the practice of trademark law. Practitioners cannot assume the accuracy of AI tools and must exercise diligence in reviewing submissions and ensure they comply with all legal requirements.
Adapting Trademark Law and Procedures
The U.S. trademark system must evolve to incorporate the reality of, and concerns associated with AI in trademark practice. The USPTO has already shown interest in embracing this new reality by implementing AI tools for applicants including introducing the AI-based Virtual Assistant on the USPTO website. However, continued guidance and future regulations will need to carefully consider AI use along with the need for human oversight. As AI continues to shape the legal landscape, clear guidelines for AI-generated materials, ongoing evaluation of AI systems for fairness and accuracy, and adjustments in processes and procedures of the trademark system will be essential to ensuring that the system continues to protect businesses and consumers.
