OpenAI Trademark Victory Sets Roadmap for Avoiding IP Missteps

OpenAI Inc. recently defeated Open Artificial Intelligence Inc. and its founder, Guy Ravine in a trademark battle that reminded the tech world that timing isn’t necessarily everything when it comes to prevailing in intellectual property battles.

As the acceleration of technological innovation speeds up—particularly in the artificial intelligence space—this case serves as a reminder to companies that despite the rapid technological advancements, they can’t ignore existing intellectual property laws or skirt legal ethics lines to try to defeat competitors.

The court’s decision provides a helpful framework for entrepreneurs seeking to protect their company names under trademark law, including a discussion of what constitutes proper use of a trademark.

Ravine started working on OAI shortly before OpenAI was founded, aiming to create a tool to allow collaboration on new AI developments. Ravine purchased the open.ai domain name in March 2015, several months before OpenAI was founded in December 2015. He didn’t file a trademark application at that time.

One day after the now well-known OpenAI was founded, Ravine applied to register OpenAI with the US Patent and Trademark Office. The USPTO rejected Ravine’s application on two bases—the mark was merely descriptive, and Ravine hadn’t demonstrated use in commerce. In response, shortly before the deadline, Ravine filed a substitute specimen—an image of the home screen of his product “Hub”—and the mark registered on the supplemental register.

As time passed, OpenAI continued to innovate and launched several successful products, including ChatGPT. OpenAI attempted to register its OpenAI mark with the USPTO but was unable to do so due to the merely descriptive nature of the mark and Ravine’s opposition to such registration.

Following ChatGPT’s success, Ravine launched similar products in his open.ai domain. This prompted OpenAI to bring this lawsuit in 2023 to settle the parties’ competing assertions of trademark rights. Most recently, OpenAI filed a motion for summary judgment, which was granted in full by the US District Court for the Northern District of California.

OpenAI also moved to cancel Ravine’s registration on the Supplemental Register, alleging he knowingly intended to mislead the USPTO. Ravine argued there were genuine disputes of material fact as to whether he intended to deceive the USPTO, further asserting the bizarre argument that attorneys regularly submit false information to the USPTO. The court granted summary judgment to OpenAI on its cancellation claims.

Although it would seem obvious that all representations and filings with the USPTO must be accurate and truthful, it may be tempting for companies to falsify or even enhance specimens to demonstrate actual use in commerce and obtain or maintain a trademark registration.

It may be enticing for companies to take shortcuts. However, this decision serves as an important reminder to companies that making false representations to the USPTO can lead to cancellation of a registration. Although such practices may result in short-term gains, they may also result in loss of the rights the company seeks to protect.

In addition to the cancellation claim, both parties claimed trademark infringement against the other. The court granted summary judgment to OpenAI on all claims. Ravine argued that as the first user of the mark, he couldn’t be liable for trademark infringement. The court, however, pointed out that a first user’s use of a mark isn’t protected beyond its original line of use “into an area occupied by an ‘intervening junior user’ who has established its own rights in the mark.”

After OpenAI became well known and acquired secondary meaning, Ravine “drifted into OpenAI’s market segment” and began offering similar products to OpenAI. The court decided that this amounted to trademark infringement.

Even though Ravine purchased the open.ai domain name prior to the launch of OpenAI, and he was using the mark in some capacity prior to OpenAI, those rights aren’t without limit, and the subsequent copying of OpenAI by using the mark for comparable goods or services was sufficient to amount to trademark infringement.

This case provides helpful information for developing companies seeking to use and protect their trademarks. It is particularly important for companies to understand the requirements when it comes to establishing use of a trademark, as well as appreciating the extent and limitation of any rights in each trademark.

For company brand management and protection, entrepreneurs in these fast-moving times of innovation must tread carefully to avoid triggering costly legal battles.

When first selecting a trademark, companies should ensure there are no existing overlapping registrations or common law usages that may impede the ability to protect the trademark in the future.

Moreover, companies should understand that prior use of a trademark doesn’t automatically extend protection to all future uses of that trademark. Rather, trademarks are protected in connection with certain goods or services, and if a senior user’s later use of a trademark extends beyond that initial protected use and any natural zone of expansion, then use of the trademark may not be protected in the same way.

The case is OpenAI Inc. v. Open Artificial Intelligence Inc. , N.D. Cal., No. 23-cv-03918, 7/21/25.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Mary Grieco is a partner and the chair of the intellectual property law practice and co-chair of the brand management and protection practice at Olshan Frome Wolosky.

Morgan Spina is an associate at Olshan Frome Wolosky and a member of the firm’s brand management and protection practice.

Write for Us: Author Guidelines