How AI Tools Can Help Brand Owners—Now and in the Future
“In theory.. if the AI algorithms can truly be trained on large datasets of existing company names, it could also be used to rule out suggested names that are already in use, given access to the right data sets.”
One of the initial challenges a new business may face is selecting a brand name that is memorable and unique, as it is the first impression consumers have of the product or service. Once a brand name is chosen, ensuring that the domain name or social media accounts for that brand are available can create additional hurdles. The feeling that every name is already taken or that one would have to get creative with the spelling in order to register across social platforms is not uncommon.
While “BRAND_NAM3_Official” may be available across the desired social platforms, further challenges can arise when applying for Trademark protections. Names, or marks, that are confusingly similar to another mark, merely descriptive or misdescriptive of a good or service, or fails to function as a source identifier are just a few ways the United States Patent and Trademark Office (USPTO) may reject an application for trademark registration and the protections that come with it. To complicate matters further, other businesses who did not register their mark may still receive common law rights based on prior use in commerce. These common law rights can be used by prior rights holders to sue subsequent users for trademark infringement or geographically limit their use of the trademark in commerce. The due diligence required in ensuring a desired brand name does not infringe another’s trademark rights ultimately requires time and a level of sophistication to parse through the entirety of the internet and other relevant databases such as the USPTO’s federal registry.
Resources such as marketing and branding agencies can help new businesses struggling to find a unique brand name, and lawyers specializing in trademark law can help evaluate infringement and refusal risks and assist in registering a trademark. That said, the costs associated with each of these options are not insignificant, leading some smaller brands and startups to choose a name, do a quick internet search for prior use, and just hope they don’t face repercussions. One of the many issues with this is that as a new business grows and becomes more recognizable, the increased exposure makes it easier for someone with pre-existing rights to find and enforce those rights against the new business. This could include financial penalties and a complete rebrand of the business causing confusion to established customers, which is why choosing and trademarking one’s brand name early is vital.
ChatGPT Tries Branding
Insert Artificial Intelligence (AI). In recent months, OpenAI’s ChatGPT, Microsoft’s Bing AI, and Google’s Bard, to name a few, are bringing the wonder of AI to the general public. While this technology is still developing, it has the potential to make tasks like inventing unique brand names more manageable. AI software such as ChatGPT can respond almost instantaneously in full sentences making users feel as if they are conversing with a knowledgeable individual. AI technology, however, is not always accurate (despite its confident tone of writing), so it is important to understand how the AI formulates its responses so users can better guide the technology. To begin, I asked ChatGPT how it could assist me in creating a brand name, and in its own words it replied:
“There are several ways that AI can be used for naming a company, including natural language processing (NLP) and machine learning (ML). NLP can analyze language patterns, identify keywords and topics, and generate names based on these insights. ML algorithms can be trained on large datasets of existing company names and brand identities to develop an understanding of naming conventions and trends.
To begin the process of using AI to generate a company name, it is important to define the brand identity and values that the name should represent. This can include factors such as the industry, target audience, and overall tone of the brand. This information can be used to train the AI algorithm and generate a list of potential names.
One way to use AI to generate a company name is to input keywords related to the brand identity and values. The algorithm can then use NLP to analyze these keywords and generate names based on related terms and concepts. For example, if the brand identity is focused on sustainability and eco-friendliness, the algorithm might generate names like “GreenPath,” “EcoWise,” or “Sustainify.”
Another approach is to use ML algorithms to analyze existing company names and identify patterns and trends. This can include analyzing the length and structure of names, the use of specific letters or sounds, and the overall tone and style of the name. By analyzing these patterns, the algorithm can generate new names that follow similar conventions and are likely to be well-received by consumers.”
Those with trademark experience may have read ChatGPT’s explanation with a “confusingly similar” red flag since the technology suggests brand names based on ones already in use. In theory, however, if the AI algorithms can truly be trained on large datasets of existing company names, it could also be used to rule out suggested names that are already in use, given access to the right data sets.
As it stands today, AI technology can still be helpful to overcome the proverbial “writers block” when trying to come up with a unique brand name. When prompted with “what are creative names for a company” ChatGPT was able to generate a list of names such as “NexusWorks,” “BrightMind,” and “StellarCore” (all of which an internet search showed are already in use). ChatGPT can also take into consideration keywords related to the brand identity and values. By responding to the AI with more details such as, “my company sells clothing” and “the average age of my consumer is 25 years old,” ChatGPT was able to return a list of names such as “Fresh Threads,” “Trendy Tribe,” and “Urban Chic” (again, all of which happened to be in use).
When I informed the AI that all the names it had suggested were already in use, it “apologized” and regenerated another list of relatively generic names combining words related to “youth” and “clothing.” The more information about my target customer and what products I sold, however, the more tailored the AI results became. For example, telling the AI the target audience was males looking for formal wear, the generated list evolved into names such as “Polished Pinstripes,” “Dapper Duds Co.,” and “Majestic Menswear.” While the somewhat random generation of names is a bit like playing Russian roulette (you never know which name might infringe upon pre-existing rights) it can help business owners get an idea of what words may resonate with their targeted audience.
The Future: AI for Trademark Search
In addition, while the AI (or at least ChatGPT) did not have the capabilities to search the internet for prior use, the analytical “skills” and recommendations the AI present are very promising. For example, at one-point ChatGPT suggested I do a trademark search. I asked the AI to do it for me, but I was told ChatGPT does not have direct access to the USPTO database, and instead gave me step by step instructions on how I could conduct my own search, and provided a link to the USPTO website. The USPTO’s Trademark Electronic Search System (TESS) can be searched in detail using various field codes. ChatGPT gave me some suggestions (when prompted) on how to use the field codes by telling me, for example clothing would be covered under International Class 025, and that I could use the field code [FW] to search full marks in TESS. While ChatGPT seemed unable to write out the actual code for me (example: “(Polished Pinstripes) [FW] and (025) [IC]”) it was able to write code in various other programming languages such as python, java, and C/C++. As the AI becomes more capable, it is not unforeseeable that given access to the USPTO database it would be able to write its own field codes to perform comprehensive trademark searches before recommending unique brand names to new businesses. The AI could then, in theory, be trained to simultaneously remove any suggestions that appear in internet searches with metatags related to the class of goods or services the mark will be used in connection with.
While the processing power associated with instantaneously performing comprehensive conflict checks across the entirety of the internet may seem like a pipe dream, AI has the potential to revolutionize the way businesses approach branding and intellectual property protection. It’s important to note that AI technology is still developing, and its suggestions are not always foolproof, but it shows promise of providing valuable resources to businesses and individuals alike. As AI technology stands today, users should still conduct their own due diligence before relying on the AI’s responses. Thankfully, the AI recognizes this and can help direct users with where to start, such as the USPTO’s website or providing the American Bar Association’s website for attorney referrals.
Image Source: Deposit Photos
Image ID: 7252917
Author: nmcandre
