By Steve Levy
Two recent UDRP decisions involving similar facts and arguments offer a study in contrasts and highlight the need for an appeals level in the UDRP as proposed at https://udrp.group and the more recent WIPO/ICA review of the UDRP. Both cases were brought by Ubiquity Global Services, owner of the AIGENT trademark, and in both, the respondents advanced nearly identical explanations for their domain registrations. Each argued that “aigent” is a contraction of “AI agent” and is a term widely used by other companies, each described their own AI-related businesses as unrelated to Ubiquity’s call-center technology, and each insisted that they had never heard of the Complainant before receiving the UDRP complaint. Despite these parallels, the two panelists reached opposite conclusions.
In the aigenthelp.com case, the Respondent submitted evidence that he operated a Texas-based company called AIgent, LLC, and argued that he had registered the domain because it matched his business name and reflected the nature of his AI-focused work which is different from that provided by the Complainant. He submitted a sworn declaration stating that he had no knowledge of the Complainant or its trademark at the time of registration. He also emphasized that “aigent” is a term used by many unrelated companies in the AI space. Nevertheless, the panel was unconvinced. It viewed the Respondent’s AI related website as “indistinguishable” from Ubiquity’s and concluded that the addition of the word “help” suggested an official support portal, increasing the likelihood of confusion. The panel ultimately found bad faith registration and use, ordering the domain to be transferred to the Complainant.
The aigentapp.com decision, issued only weeks later, took a very different approach. Here, the Respondent—also based in Texas—explained that he was developing a real-estate-focused “Agent App” and had chosen the domain because it fit the project’s branding. Like the first Respondent, he argued that “aigent” is a commonly-used blend of “AI” and “agent” and that he had never heard of the Complainant before registering the domain. This panelist accepted the explanation as credible and emphasized the breadth of third-party use of the term. Finding no evidence that the Respondent had targeted the Complainant’s mark or attempted to trade on its reputation, the panel concluded that he had rights and legitimate interests in the domain and denied the complaint.
What makes these cases so striking is not simply that the Respondents told similar stories, but that the panelists interpreted those stories in such dramatically different ways. One saw imitation and bad faith targeting whereas the other saw a common term used for different AI related services with no sign of targeting. When two cases involving the same Complainant, the same trademark, and nearly identical Respondent arguments can produce such divergent outcomes, it highlights a structural issue in the UDRP. After twenty-five years of developing consensus principles, the system still lacks an appeals mechanism to correct outlier decisions. The proposals made by the UDRP.group and the WIPO / ICA Report suggest that a losing party in a single-member panel case have the opportunity to appeal the decision to a three-member team of highly experienced UDRP panelists. While this would not create “precedent”, as the term is used in courtroom litigation, it would serve to correct wrongfully decided cases and serve as authoritative guidance for future cases.
As such proposals move very slowly, it’s not anticipated that we’ll see an appeals level anytime soon. However, it’s an idea worthy of serious consideration and debate and hopefully it will be implemented at some point to keep UDRP outcomes within the guardrails of the consensus that has now been built up in more than a quarter of a century and over 100,000 UDRP decisions.