On March 2, 2026, the U.S. Supreme Court declined to hear Thaler v. Perlmutter, letting stand the D.C. Circuit’s ruling that copyright protection under federal law requires a human author. For businesses using AI tools to create logos, write marketing copy, or develop software, that decision isn’t abstract. It directly affects whether what your business creates with AI is yours to protect and whether a competitor who copies it can be stopped.
The businesses asking these questions right now aren’t doing anything unusual. They’re using ChatGPT for website content, Midjourney for brand visuals, or GitHub Copilot for internal tools. The assumption is usually the same: we paid for the tool, we gave it the prompt, so we own the output. That assumption deserves a closer look before it becomes a business problem.
Michele L. Ross founded M. Ross & Associates, LLC after practicing at Paul, Hastings, Janofsky & Walker and Carter, Ledyard & Milburn in New York City. We advise small and mid-sized New Jersey businesses on IP registration, licensing, and the commercial contracts that govern these relationships. What follows draws on that work to address the ownership and protection questions AI has made urgent for businesses of every size.
Using AI in your business? Protect your intellectual property with M. Ross & Associates, LLC. Call (201) 897-4942 today to safeguard your innovations.
What the Supreme Court’s Refusal to Hear Thaler v. Perlmutter Means for AI-Generated Content
Stephen Thaler is a computer scientist who created an AI system he named the Creativity Machine. He submitted artwork it generated, titled “A Recent Entrance to Paradise,” to the U.S. Copyright Office and listed the AI as the sole author. The Copyright Office refused registration. The D.C. Circuit agreed, holding that the Copyright Act requires human authorship, and the Supreme Court’s March 2026 refusal to hear the case means that ruling stands without further review.
The Copyright Office has registered AI-assisted works, and that distinction matters. Where a human exercised meaningful creative control over the final product, selecting, arranging, and refining what the AI produced through genuine artistic judgment, registration is available for that human’s contribution. The line isn’t whether AI was involved. The line is whether a person made real creative decisions, not just typed a prompt and accepted whatever came out.
For a business, this means a logo generated entirely by an AI system with no human creative shaping may carry no copyright protection at all. Anyone who copies it may face no infringement claim. That’s a meaningful gap if the business has been treating that logo as a protected asset.
Trademarks & Trade Secrets: Two Different Risks When AI Touches Your Brand
Copyright and trademark are separate legal tools that protect different things. A business whose AI-generated logo lacks copyright protection isn’t necessarily without any recourse. Trademark rights arise from using a mark in commerce to identify the source of goods or services, not from registering a copyright. If your business has been using that logo in the market long enough to build recognition, a competitor who adopts a confusingly similar mark may still face trademark infringement and unfair competition claims, even if the underlying design never qualifies for copyright registration. That said, trademark protection covers commercial identity. It doesn’t make your creative assets copyable with impunity everywhere else. A business relying on an AI-generated design needs to think clearly about which protection it actually has and which it doesn’t, because the two aren’t interchangeable.
Trade secret law protects confidential business information, including formulas, processes, client lists, and pricing strategies, as long as the owner takes reasonable steps to keep it secret. Entering that information into a publicly accessible AI platform destroys that protection. Once shared with an external system whose data handling and training practices may expose it to third parties, the information is no longer confidential. Trade secret misappropriation claims depend on secrecy that either never existed or that the business itself eliminated.
Patents & AI-Assisted Inventions
The patent side of this question was settled earlier. In Thaler v. Vidal, the Federal Circuit ruled that only natural persons can be named as inventors under the Patent Act. The Supreme Court declined to review that decision in 2023. The USPTO’s revised inventorship guidance, issued in November 2025, reaffirms the same principle: AI can’t be named as an inventor and is treated as a research tool, comparable to laboratory equipment or specialized software.
Patent inventorship still turns on conception, specifically the formation of a definite and permanent idea of the complete and operative invention. A human inventor who uses AI to model possibilities, accelerate testing, or generate candidate structures needs to articulate the specific idea they personally formed. Pointing to a general goal given to the AI system isn’t enough. For New Jersey businesses in life sciences, technology, or advanced manufacturing, where AI-assisted R&D has become common, that makes documentation practices a practical necessity, not just a legal formality.
Where Ownership Actually Gets Decided: Contracts & Vendor Terms
The cases above address what the law protects. Contracts often determine who owns it. These two things can point in different directions, and in commercial disputes, the contracts usually win.
AI Platform Terms of Service
Most AI platforms include terms that grant them rights to user inputs and, in some configurations, to outputs as well. Before entering proprietary product specifications, client data, draft contracts, or any confidential material into an AI tool, review what those platform terms actually say about data use, training, and ownership of generated content. “Terms of service” sounds routine. For a business with valuable trade secrets or in-development IP, it isn’t.
Vendor & Freelancer Agreements
If a contractor uses AI to produce deliverables for your business, does the agreement require them to disclose that? Does it include representations about originality and ownership of what they deliver? Many standard freelancer agreements don’t address this because they predate the widespread use of generative AI. Intellectual property assignment clauses in those agreements often assume the contractor created the work themselves, which means the chain of title for AI-assisted deliverables may be unclear from the start.
NDAs That Predate Generative AI
Confidentiality provisions in agreements drafted before 2022 or 2023 typically say nothing about whether an employee or counterparty may input protected information into an external AI system. An NDA that prohibits sharing trade secrets with competitors but doesn’t address AI tools may not cover exactly the disclosure scenario a business is most concerned about today.
Practical Steps to Protect Your Business’s IP When Using AI
These aren’t universal requirements, but they’re grounded in how copyright, patent, and trade secret law currently treats AI involvement. For most New Jersey small and family-owned businesses, they represent the difference between having a defensible IP position and discovering too late that you don’t.
- Document the human contribution. For any AI-assisted content or invention you intend to protect, keep records of the prompts used, intermediate drafts, and the specific edits or refinements a person made. This creates a paper trail showing meaningful human creative or inventive input, something both Copyright Office registration and patent applications for AI-assisted work depend on.
- Build a written AI usage policy. An AI usage governance policy identifies which tools employees may use for company work, which categories of information may never be entered into any AI system (trade secrets, client data, pending patent disclosures), and what disclosure is required when AI is used to create deliverables. It’s also what makes trade secret protection credible; courts look at whether a business took reasonable steps to maintain secrecy.
- Update your contracts before the next project starts. Vendor agreements, freelancer agreements, and NDAs should be reviewed now for the gaps described above. That means adding AI disclosure and warranty provisions to work-for-hire agreements, tightening confidentiality provisions to cover AI platform inputs, and confirming that IP assignment language captures AI-assisted work clearly.
- Review platform terms before inputting proprietary material. This takes an hour and can prevent a problem that’s very difficult to undo. Know what rights an AI platform claims over inputs and outputs before your team uses it for anything sensitive.
- Name the human creator clearly in filings. When filing for copyright registration or a patent application involving AI-assisted work, identify the human author or inventor and be prepared to describe their specific contribution in concrete terms. “We used AI as a tool” isn’t sufficient; the filing should reflect what the person actually decided or created.
The rules around AI and intellectual property are still developing, nationally and here in New Jersey. The Supreme Court’s Thaler v. Perlmutter decision settled one important question but left others open, and the USPTO continues to issue guidance as cases and technology evolve. This is precisely the kind of area where having outside general counsel monitoring developments on an ongoing basis, rather than waiting for a dispute to surface, protects the business in ways that reactive legal work can’t replicate.
We work with New Jersey businesses on IP registration, licensing, vendor contract review, and the contract-level planning that keeps ownership questions clear before they become disputes.
If your business is using AI tools and you’re not sure whether your IP position is solid, we can help you find out. Reach us at (201) 897-4942.