/>iMeanwhile, the US Patent and Trademark Office (USPTO) has issued refusals to register Google’s federal trademark applications for GEMINI (US Application Ser. Nos. 98202646 and 98418146) based on an alleged likelihood of confusion with Gemini Data’s registered marks.
Prior to publicly rebranding its line of AI products from BARD to GEMINI in February, Google filed a federal trademark application seeking to register the GEMINI mark in connection with certain AI-related software services, followed by another application covering additional AI-related software products shortly after the announcement. In May and August, respectively, the USPTO refused registration based, in part, on an alleged likelihood of confusion with Gemini Data’s earlier registrations for stylized GEMINI and GEMINI DATA marks, which cover data collection and analytics software that uses AI and machine learning technologies. Thereafter, Gemini Data filed its complaint in federal court, asserting that Google had adopted the GEMINI mark despite actual knowledge of Gemini Data’s earlier marks, having “undoubtedly” conducted a trademark clearance search prior to rebranding. The complaint also claims Google attempted to acquire Gemini Data’s rights through an anonymous strawman buyer, among other claims alleging willful infringement.
Despite the serious allegations, the complaint includes a seemingly lighthearted footnote quoting a purportedly incriminating admission by Google’s AI chatbot: “In an almost laughable example of Google’s hubris, if you ask Google’s Gemini application if it is aware that it is infringing upon the trademark of Gemini Data Inc., it responds ‘Yes’ and that ‘[i]t’s a developing situation.’” Of course, the admissibility of such a statement is a matter for another day.
Google has yet to file an answer to the complaint, though it has responded to the USPTO’s office action refusals by requesting a suspension of examination pending the disposition of the related court action.
More Than Meets the Eye
On the surface, Gemini Data’s claims, including alleged rights arising from use well prior to Google’s public launch of its GEMINI-branded AI products earlier this year, seem to present a straightforward case of trademark infringement. However, a deeper dive suggests this may not necessarily be the case. For instance, a review of the USPTO’s records reveals that Gemini Data’s own trademark registrations had been refused based on an alleged likelihood of confusion with other GEMINI-formative marks owned by unrelated third parties, and that Gemini Data had restricted the scope of its goods and services and submitted detailed arguments in an effort to persuade the USPTO to withdraw the refusals. These arguments included, among other things, claims that the GEMINI mark is weak, dilute, and entitled to only a narrow scope of protection by virtue of widespread third-party use and registration of GEMINI-formative marks in connection with similar or related goods or services. Gemini Data also emphasized the “highly stylized” nature of its marks, arguing that this served to “immediately distinguish” them from the cited GEMINI-formative marks, as well as specific and nuanced differences in the nature and purpose of its goods and services. While these statements made during the USPTO application process do not necessarily prevent Gemini Data from taking contradictory positions in its pending litigation against Google, they will likely be cited by Google in an effort to build a defense against Gemini Data’s claims of infringement.
Takeaways
While still in its early stages, this case highlights the importance of trademark clearance searches and careful consideration of potential conflicts before adopting new brands, especially in the context of emerging technologies such as AI. Additionally, brand owners should be mindful of statements made on the record during the trademark application process, as they could later be cited against the brand owner and undermine potential enforcement efforts, ultimately proving to be the quintessential “double-edged sword.” The outcome of this case may also have broader implications with respect to the scope of trademark rights afforded to marks used in connection with AI-related goods and services, including, for example, the extent to which nuanced differences in the purpose or functionality of certain AI technologies — or a given brand owner’s target consumers or trade channels — can mitigate against a likelihood of consumer confusion, given the increasing ubiquity of AI tools in virtually all industries. Brand owners should closely monitor developments in this area and be prepared to adapt their strategies to safeguard their rights.
