OAKLAND, Calif. (CN) — A federal judge on Thursday dismissed individual counterclaims against two cofounders of OpenAI but opted to keep a trademark infringement accusation against the ChatGPT company intact.
The ruling by U.S. District Judge Yvonne Gonzalez Rogers stemmed from a suit filed by OpenAI against Open Artificial Intelligence Inc., which it’s suing over its trademark and name. Open Artificial Intelligence, in turn, filed a counter-complaint arguing trademark infringement.
Rogers’ ruling in her motion to dismiss focused on the trademark infringement counterclaim, which argued that consumer confusion occurred when OpenAI’s ChatGPT tool was announced on Nov. 30, 2022.
Guy Ravine, founder of Open Artificial Intelligence, has argued that he created the idea of OpenAI in 2014 after speaking to industry leaders. He claimed Sam Altman and OpenAI cofounders stole his foundational ideas and used his domain after he refused to sell rights to the company.
Open Artificial Intelligence had a website URL of open.ai, which saw a 1,100% increase in traffic after the ChatGPT announcement. It regularly gets emails from people who think it’s linked to OpenAI. Additionally, people who visited open.ai in 2023 typed “chatgpt” or a variation into its AI tool over 6,500 times.
OpenAI has claimed that it launched an image generator tool in September 2022. Ravine launched a similar tool two months later. It was only then that competing products existed, and the counterclaim makes no accusation of consumer confusion at that point.
If the judge allowed a claim based on the Nov. 30, 2022, rollout of ChatGPT, then anyone could pick “their own convenient day” when determining when infringement began, even when the companies already had competing products, OpenAI argued.
“Plaintiff construes defendants’ counterclaims much too narrowly,” Rogers wrote. “The allegations are based on a theory of a company founded years prior to the launch of ChatGPT, and whose use of the mark dates back to 2015.”
OpenAI also argued that there was no mental association by a large segment of consumers between Ravine’s mark and its own. Ravine countered that the frame of reference shouldn’t be the entire consuming public, but instead the intended audience — in this case, the niche artificial intelligence community.
The judge agreed with Ravine, ruling that he began using the disputed mark in 2015 and kept using it continuously. He also had a following of users within a growing industry. The emergence of OpenAI and a competing mark disrupted that following.
Two elements are needed to argue trademark infringement, Rogers wrote — secondary meaning and likelihood of confusion. A jury should decide secondary meaning. Ravine provided proper accusations of likelihood of confusion in his amended counterclaim, meaning that portion of it can proceed.
“Whether defendants can actually garner evidence of the same is a separate issue,” Rogers added.
However, the judge dismissed individual claims against Sam Altman and Greg Brockman, cofounders of OpenAI.
Specific action by a corporate officer is needed to hold them personally liable for trademark infringement, Rogers wrote. Ravine claimed that the men knew of his venture and “rushed” to advertise their own business using the mark.
Ravine also argued they had a meeting in 2015, when Altman asked Ravine to change his business’ name. Additionally, Ravine claimed that in 2022, Altman asked about buying the open.ai domain and related intellectual property rights.
Rogers ruled that more was needed to hold Altman and Brockman personally liable.
“As an initial matter, statements made ‘on information and belief’ without additional facts are too conclusory to credit even at this early stage,” the judge wrote. “Second, the general statement that Altman is influential is likewise insufficient to demonstrate affirmative participation in the alleged infringement. In short, the court is left with allegations that Altman received an email and met with Ravine once, and that he later expressed interest in purchasing Ravine’s IP rights.”