It is understandable why OpenAI has been cautious when it comes to their brand recently. A few of the businesses that have applied for trademarks to the United States Patent and Trademark Office recently include ThreatGPT, MedicalGPT, DateGPT, and DirtyGPT.
Everyone is capitalizing on the incredible success of ChatGPT, the conversational AI which OpenAI launched in November. This chatbot has been developed using OpenAI’s deep learning technology, with GPT-4, their newest version, being released just a month ago.
It is unsurprising that OpenAI quickly requested a fast-tracking procedure for their “Generative Pre-trained Transformer” (GPT) trademark confirmation, due to the numerous imitations and bogus applications that had started to surface.
Last week, OpenAI’s request was not accepted. The agency declared that OpenAI’s lawyers had ignored paying the accompanying charge and furnishing the “suitable written proof supporting the justification of special action.”
It could take as long as five months for OpenAI to reach a conclusion, according to Jefferson Scher who is a partner in the intellectual property group of Carr & Ferrell and chair of the firm’s trademark practice group. However, even then, there is no assurance that the ultimate result is known, explains Scher.
Scher was asked if OpenAI might encounter resistance due to the “T” in GPT standing for “Transformer,” which was first unveiled by researchers at Google in 2017 and is now commonly utilized. Scher posed the question, “Can GPT still be a brand even if it derives from a very illustrative source?” He highlights IBM as an example of a company that had a weakly descriptive name from the beginning, which suggests there’s no promise OpenAI will own GPT, according to Scher. However, this gives some hope.
Scher notes that OpenAI has had years of experience working with “GPT”, having already released its original Generative Pre-trained Transformer (GPT-1) model in October of 2018.
Scher commented on the oddity of the situation, that usually brand recognition is achieved with prolonged use in the marketplace, yet OpenAI was a relatively unknown to AI researchers until they released their deep learning model generation, DALL-E 2, and ChatGPT, leading them to become an instant success.
Even if the USPTO examiner has no reservations about OpenAI’s request, it will then be brought to a opposition phase, where other entities in the market can give reasons why the ‘GPT’ brand mark should be refused.
Scher outlines that OpenAI’s point could be contested, wherein a rival would question whether the acronym “GPT” is actually the property of OpenAI and instead portrays generative AI in general.
One way of asking the United States Patent and Trademark Office (USPTO) for a ruling on public opinion is by conducting a survey. According to Scher, this could involve asking a sample of random American citizens a question, yet this is an expensive endeavour that the USPTO is not likely to pay for. Either the challenger or OpenAI, or both, would need to finance this project.
The way in which “GPT” has been mentioned in public, from comedians on late-night programs to articles written in the media, can also help form public opinion. According to Scher, if people are not handling the name as if it were exclusive, then ultimately, a court case would decide whether or not it is trademarkable.
The lengthy process this would require is definitely not something that OpenAI is keen on.
You have to wonder why the company waited so long to safeguard “GPT”. According to Scher, they may have been stunned by the success of the product. They have tried to get ahead of the situation in China, where they have not yet released ChatGPT and may not be able to, as they looked to apply a trademark related to it.
In Scher’s view, it is no longer just some random letters: GPT. If a new business were to ask him if it is safe to use it, his answer would be no.
Scher states that OpenAI may potentially become so well-known that their prominence will affect their ability to trademark. For example, Rolex’s fame has given it protection beyond its field, so that its trademark cannot be used for anything else. This same possibility exists for OpenAI, given that they can successfully demonstrate that “GPT” is a unique identifier for the organization. The company’s well-known trademark can be safeguarded against unauthorized use widely, even if tracking down violators would be costly.
It may be beneficial for the company throughout their period of stagnation. With increased passing time, OpenAI will likely gain more users as well as media attention, which would make their desired result much more likely to occur.
Scher suggests that OpenAI may be becoming a more familiar name within the average household.