AdobeStock 704394220
November 19, 2026

AI Helped Create It. Can Your Business Protect It?

A product team uses AI to work through a design problem. Marketing uses it to develop campaign material. An employee pastes a technical specification into a chatbot to get a clearer summary. Each use may save time, and each raises a different intellectual property question.

For the design, the question is who invented it. For the campaign material, it is what a human authored. For the technical specification, it is whether the company kept confidential information confidential. These questions are easiest to answer while the work is underway, and they become harder to determine when a patent application, copyright registration, or dispute requires someone to reconstruct what happened months or years later.

An AI-assisted invention still needs a human inventor

Using AI during development does not by itself prevent an invention from being patented. An engineer might use a model to explore alternatives, analyze test results, or help solve a technical problem. The question remains whether a natural person conceived the invention claimed in the patent application, as the AI system cannot be named as an inventor.

The U.S. Patent and Trademark Office revised its guidance on AI-assisted inventions in November 2025. It applies the same inventorship standard whether or not AI was used. That does not make every AI-assisted design easy to assess; a team still needs to identify the human contribution to the claimed invention.

Consider an engineer who defines the technical problem, proposes a design, and uses AI to test alternatives. That presents a different record from one in which a model proposes the operative design and the team simply accepts it. Neither description alone resolves inventorship; the details of conception and the eventual patent claims matter. But the distinction shows why “we used AI” is not a useful answer to “who invented this?”

Businesses developing patentable technology should keep the ordinary development record: who proposed the solution, how the design changed, what the AI tool contributed and how it was used, and who made the decisions that produced the final invention. Prompts and outputs may help establish that history, but they do not substitute for a human inventor.

Copyright asks a different question

Copyright protects human expression. Under the U.S. Copyright Office’s current approach, using AI as a creative tool does not prevent protection for what a person actually authored, but copyright does not extend to material generated by AI without sufficient human authorship.

That distinction matters to a business commissioning a logo, building a website, or producing a campaign. If someone enters a prompt and selects an image from the results, the company should not assume it can claim copyright in the image simply because the prompt took effort or the company paid for the tool. The Copyright Office does not treat prompts alone as sufficient authorship of the resulting expression.

However, a person may contribute protectable expression by creatively modifying generated material or selecting and arranging it as part of a larger work. Protection then extends to the human-authored contribution, not automatically to every element in the finished file. The precise line depends on what the person did.

For a commercially important asset, that is a practical design question as much as a registration question. Who created its key visual or written elements? Which parts came from a model? What did a person change, combine, or add? If the business expects to prevent competitors from copying the asset, it should understand the answers before making the asset central to its brand or product.

A platform’s permission to use an output commercially does not answer whether the output is copyrightable. Those are separate questions. If the business later seeks registration, it must accurately identify its human-authored contribution and disclose AI-generated material as required by Copyright Office guidance.

Keep a record of the work, not just the finished result

Patent inventorship and copyright authorship involve different legal tests. They share a practical difficulty: the finished design or image rarely tells the whole story of how it was made.

A useful record need not preserve every abandoned prompt. It should let the business identify the tool used, the relevant inputs and outputs, the people involved, and the decisions that shaped the result. That record can support an accurate patent or copyright assessment, as it cannot create a human contribution that never occurred.

Confidentiality can be lost during the same workflow

The information entered into an AI tool may be more valuable than its output. Specifications, source code, process settings, customer information, and unreleased plans can all find their way into prompts or uploaded files. AI meeting assistants may collect comparable information without anyone pasting it into a chat window.

Trade secret protection depends in part on reasonable measures to keep valuable information secret. Submitting it to a provider without appropriate confidentiality protections can jeopardize that position. The result is fact-specific. It depends on what was shared, which tool and account were used, the applicable terms and settings, and who could access or retain the information. A consumer account and an enterprise account for the same product may present different answers.

A workable company policy should therefore say which tools and accounts are approved, what information employees may enter, and when they need review before using a tool. The business should also know whether the provider has confidentiality obligations and what its terms say about access, retention, and use of submitted information. An instruction to “be careful with AI” leaves those decisions to each employee.

AI can be useful in creating valuable work. The business question is whether the way it uses AI preserves the rights it expects to have. Identify the human contribution, keep a record of it, and know what information is leaving the company before the tool becomes part of the routine.

Marcus DiBuduo is a Director at Fennemore, supporting entrepreneurs of all sizes with strategic counsel on all aspects of intellectual property law, including the acquisition, enforcement, and licensing of patents, trademarks, copyrights, and trade secrets. He can be reached at mdibuduo@fennemorelaw.com.