Trademark Risks In The Age of Generations AI

When I started reading about trademark law, I honestly thought the biggest threats to a brand were competitors copying names or someone selling knockoffs in a shady marketplace. That picture changed pretty quickly once I started looking at what AI tools are doing to the trademark system right now. The issue is not straightforward, and the courts are still figuring it out which, honestly, makes it a fascinating time to be studying IP law.

HOW IS AI-GENERATED CONTENT INFRINGING TRADEMARKS

Trademarks are really important for businesses to protect their identity and reputation so brands can stay unique. Not be used without permission. These days businesses, small startups are using AI tools to create brand names, logos, taglines and even entire visual identities. It is easy to see why they like these tools. They are fast and cheap. The results often look very professional. A business owner can have a brand ready in just one afternoon.

The problem is that trademark law has not changed to deal with this way of doing things. According to the Lanham Act, which’s the main federal trademark law in the United States infringement happens when a mark is likely to cause confusion among consumers about where goods and services come from. The key thing to remember here is “likely”. Courts do not need proof that anyone was actually confused.. More importantly courts do not care if the infringement was on purpose or not.

This is where AI generated content creates an issue. These AI tools are trained on amounts of existing brand content like logos, fonts, names and visual styles. When you ask an AI to generate a logo for a wellness brand it is using patterns from thousands of wellness brands it has already seen. Sometimes the result is original enough to be okay. Sometimes it is too similar, to something that already exists. The business owner has no idea that AI generated content is infringing trademarks.

WHAT THE CASES ARE STARTING TO SHOW

1. WEX Inc. v. HP Inc. (2024)

This case is actually one of my favorites to discuss because it shows how AI and trademark law are colliding in unexpected ways.

WEX Inc., a financial services company, sued HP over the use of the name “HP WEX” for a software product. WEX claimed this was likely to cause confusion with their registered mark “WEX.” What made the case interesting from an AI angle was HP’s defense, they argued that a news article submitted as evidence of consumer confusion had been “created using generative artificial intelligence,” and therefore no real person had actually been confused. 

The court rejected this argument. It held that the AI-generated article still indicated that consumers could confuse the two brands as being affiliated, especially because a human editor had reviewed the article. The court granted WEX a preliminary injunction against HP’s use of the mark.

What this tells us is the even AI-generated content an serve as evidence of likelihood of confusion. The line between” AI did it” and “a human was responsible” is not as clean as defendants might hope.

2. Andersen v. Stability AI ltd.(2024)

This case, decided by a California district platforms had incorporated a kind of “trade dress database”- essentially the ability to recall and recreate the distinctive visual styles of individual artists. The court denied Midjourney’s motion to dismiss the trade dress infringement claims, finding that the allegation was plausible enough to proceed.

Now, trade dress protection covers the overall visual appearance of a product or brand- things like the distinctive layout of a store or the look of a product’s packaging. Extending this to an artist’s recognizable visual style is a relatively new argument, but the fact that the court let it survive dismissal is significant. It suggests that courts are willing to apply traditional IP frameworks to AI-generated content rather than treating AI as some kind of legal exception. 

3. The New York Times v. OpenAI and Microsoft (ongoing)

This is probably the most talked about AI-related IP case right now, and while it is primarily a copyright dispute, it has trademark dilution claims embedded in it as well.

The Times alleged that OpenAI and Microsoft used millions of its copyrighted articles to train their AI models without permission. Among other things, the complaint includes claims of trademark dilution- the argument being that AI-generated “hallucination” falsely attributed to The Times damage its brand reputation and distinctiveness. 

The court, as of march 2025, denied the defendants’ motions to dismiss the trademark dilution claims, allowing them to proceed. This matters because trademark dilution-unlike straightforward infringement does not require proof of consumer confusion. It only requires that the use blurs or tarnishes the distinctiveness of a famous mark. If courts consistently allow dilution claims against AI companies for how their outputs affect established brands, the liability exposure for the entire AI Industry could be substantial.

4. Thaler v. Perlmutter (2023)

Thaler v. Perlmutter in the U.S. District Court of D.C. Affirmed the U.S. Copyright Office’s denial of copyright registration for an output generated solely by artificial intelligence in the absence of human authorship. Though the decision dealt with copyright law, it is relevant for intellectual property questions more broadly, such as trademarks. The court confirmed that humans were still at the heart of the IP system. For any businesses relying on AI to generate brand assets, this case acts as a warning that there could be significant ambiguity in rights protection where human contribution is limited. Though trademarks do not rely on authorship for their existence but on actual use in trade, there are likely to be growing concerns on the part of regulators and courts over the degree of human control used in the generation of AI-produced materials. This implies that careful human oversight will be necessary when creating, implementing or registering AI-generated names, logos, and branding material.

WHY THIS MATTERS FOR BUSINESS

The cases I mentioned earlier are about big companies and popular platforms. The same rules apply to any business that uses AI for branding.

If an AI tool creates your logo and it looks too similar to a trademark that is already registered you could be in trouble. Even if you did not know about the existing trademark. This is one of the things about trademark law. It does not matter if you meant to cause harm or not. What matters is whether people might get confused.

The question is, who actually owns a trademark created by AI? The USPTO says that a human has to create property for it to be protected. For copyright it is clear that AI-generated works with no input cannot be copyrighted.

There is still some confusion, about trademarks. This is because trademark protection is based on how you use it in business not who created it.  Trademark examiners are now looking closely at trademarks created by AI.

 SO WHAT SHOULD ACTUALLY BUSINESS DO?

They should not stay away from AI tools completely. AI tools are really helpful when it comes to branding. The problem is when businesses think that what AI tools say is the answer.

Businesses should do a trademark clearance search before they start a brand. This search is not about looking something up on Google or using a free online tool. It is about looking at the USPTO register for trademarks that’re similar to the one they want to use in the same class of goods or services.

It is an idea to hire a trademark attorney to give an opinion on whether a trademark is clear to use. This is something that a lot of business owners do not want to pay for but it is better than having to change the brand name later. Changing a brand name can be very expensive.

Filing for a trademark early is also important. In the United States the first person to use a trademark has the rights, to it. Filing an application first gives a business priority over others who file later.

If a business is building a brand around a name they should file for a trademark before they launch not after they launch. This is because trademark rights are based on who uses a trademark and who files for it first like the business that is building a brand around a name.

PREVENTIVE MEASURES FOR INDIAN BUSINESSES USING AI TOOLS

Indian businesses can reduce trademark risks by adopting a structured compliance approach. First, conduct trademark clearance searches through the Indian Trade Marks Registry before finalizing any AI-generated brand name or logo. Second, maintain records showing human involvement in selecting and modifying AI-generated outputs. Third, use professional legal review for high-value branding decisions. Fourth, establish internal AI governance policies that require verification of originality and potential conflicts with existing marks. Finally, file trademark applications at an early stage and monitor the market for potentially conflicting uses.

THOUGHTS TO CLOSE

I have to say that when I first started looking into this topic I was surprised that the courts are moving so fast. If we look at cases like WEX v. HP and Andersen v. Stability AI we can see that judges are not waiting for Congress to make laws about Artificial Intelligence and intellectual property. They are using the laws we already have and trying to figure out if they work for Artificial Intelligence.

Sometimes the laws work well for Artificial Intelligence. Sometimes they do not fit well. It is clear where things are going.

Artificial Intelligence is not something that can protect you from getting in trouble for using someone Trademark. If Artificial Intelligence makes anything easier to use someone Trademark by mistake and it is harder to know what will happen when you do. For anyone who is building a brand now this is probably the most important thing to know about Artificial Intelligence. The law says that not knowing something is not an excuse and this is especially true, for Artificial Intelligence.

REFERENCES

Lanham Act, 15 U.S.C. §§ 1051 et seq.

WEX Inc. v. HP Inc., U.S. District Court (2024).

Andersen v. Stability AI Ltd., U.S. District Court for the Northern District of California (2024).

The New York Times Co. v. OpenAI, Inc. and Microsoft Corp. (ongoing litigation).

Thaler v. Perlmutter, 687 F. Supp. 3d 140 (D.D.C. 2023).

United States Patent and Trademark Office (USPTO), Trademark Manual of Examining Procedure (TMEP).

Controller General of Patents, Designs and Trade Marks, Government of India – Trade Marks Registry guidance and resources.