Your marketing team uses ChatGPT to draft blog posts. Your design team uses Midjourney to create product images. Your engineering team uses Copilot to write code. The question you probably have not asked yet: do you actually own any of it?
The answer, as of mid-2026, is more complicated and more consequential than most businesses realize. The US Copyright Office, federal courts, and the Supreme Court have all weighed in, and the outcome is clear on one critical point: purely AI-generated works receive no copyright protection in the United States. They belong to everyone, which means they belong to no one.
If your business relies on AI tools to create content, code, images, or other creative output, you need to understand exactly where the legal lines are, what you can protect, and what you cannot.
The US Copyright Office Position on AI-Generated Works
The US Copyright Office has been consistent on this point since its initial guidance in 2023: copyright protection requires human authorship. Works generated entirely by artificial intelligence, without significant human creative input, do not qualify for copyright registration.
This means that if you type a prompt into an AI image generator and it produces an image, that image is not copyrightable. You cannot register it. You cannot sue someone who copies it. You cannot license it exclusively. The same applies to text, music, code, and any other output that an AI generates without meaningful human involvement in the creative expression itself.
The Copyright Office draws a distinction between using AI as a tool and having AI act as the author. If a human uses AI to assist in a creative process but makes significant creative decisions (selecting, arranging, editing, modifying the output), the resulting work may qualify for partial copyright protection covering the human-authored elements. The purely AI-generated portions remain unprotectable.
Thaler v. Perlmutter and the Supreme Court's Refusal to Hear the Case
The most important legal case on AI authorship is Thaler v. Perlmutter. Stephen Thaler attempted to register a copyright for a visual artwork created entirely by his AI system, DABUS. He listed the AI as the author and himself as the owner through a work-for-hire theory.
The Copyright Office refused registration. Thaler sued. The US District Court for the District of Columbia ruled against him in August 2023, holding that copyright law requires a human author. The DC Circuit Court of Appeals affirmed that ruling.
On March 2, 2026, the Supreme Court declined to hear Thaler's appeal, leaving the lower court rulings intact. This effectively settles the question at the federal level for the foreseeable future: AI cannot be an author under US copyright law. Works created entirely by AI systems have no copyright protection and enter the public domain immediately upon creation.
No other major jurisdiction has taken a different position. The European Union, the United Kingdom, China, Japan, and Australia have all declined to recognize AI as an author or inventor for intellectual property purposes. The global consensus aligns with the US position.
What Businesses Can and Cannot Protect
This is where it gets practical. Understanding the legal framework matters only if you know how it applies to your day-to-day operations.
Content you cannot protect: any text, image, code, audio, or video generated entirely by AI without significant human creative input. If you prompt an AI tool and publish the raw output without substantial editing, that output has no copyright protection. A competitor can copy it freely. A stock photo site can distribute it. Anyone can use it without your permission.
Content you may be able to protect: works where a human made significant creative decisions beyond the initial prompt. This includes selecting and arranging AI-generated elements into a larger composition, substantially editing or rewriting AI-generated text, combining AI-generated elements with original human-created elements, and using AI output as a starting point that you then transform through your own creative judgment.
The key factor is the degree of human creative control over the final expression. The Copyright Office evaluates this on a case-by-case basis. There is no bright-line rule for how much human editing is "enough." But the direction is clear: the more human creative decision-making involved in producing the final work, the stronger the copyright claim.
How to Structure Workflows to Preserve IP Rights
If your business uses AI tools in its content creation process, you should restructure your workflows to maximize the copyrightability of the final output. This does not mean stopping AI use. It means using AI in a way that preserves your ability to claim ownership.
Use AI as a drafting tool, not a publishing tool. Generate initial content with AI, then have a human writer substantially revise, restructure, and add original analysis, examples, and insights. The human-authored revisions are copyrightable. Document the process so you can demonstrate the human contribution if challenged.
For visual content, use AI-generated images as a base and have a designer modify, composite, or incorporate them into a larger original design. The original design elements and the creative arrangement are protectable. The raw AI-generated image alone is not.
For code, use AI-assisted coding tools to generate boilerplate or initial implementations, then have your developers review, modify, and integrate that code into your larger proprietary codebase. The overall architecture, the selection and arrangement of components, and any original code written by your developers are protectable.
Keep records. Save drafts showing the AI-generated starting point and the human modifications at each stage. If you ever need to enforce your copyright or defend against an infringement claim, this documentation establishes the human authorship that copyright requires.
Contract Provisions for AI-Generated Content
Contracts between businesses and their employees, contractors, and vendors need to address AI-generated content explicitly. Standard IP assignment clauses were written for a world where humans created everything. That world no longer exists.
If you hire a freelance writer or a design agency, your contract should specify whether the contractor is permitted to use AI tools in creating deliverables. If AI use is allowed, the contract should require the contractor to disclose which portions were AI-generated and which were human-authored. It should require the contractor to retain documentation of the human creative process. It should assign to you all rights in the human-authored portions (which are copyrightable) and grant you an unrestricted license to use any AI-generated portions (which are not copyrightable but may still need contractual permission depending on the AI tool's terms).
Employment agreements should include similar provisions. If your employees use AI tools in their work, your IP assignment clause should cover both traditional work product and AI-assisted work product. Define what level of AI use is permitted and what documentation is required.
Vendor and SaaS agreements increasingly include clauses about AI use as well. If a vendor uses AI to deliver services to you, the contract should address IP ownership of AI-generated deliverables, data handling (is your data used to train AI models?), and quality assurance for AI-generated output.
Terms of Service Implications for AI Output Ownership
Every AI tool your team uses has its own terms of service, and those terms affect your rights to the output. This is an area most businesses overlook entirely.
OpenAI's terms, for example, assign output ownership to the user, but they also state that other users who generate similar output from similar prompts may receive identical or near-identical content. You have no exclusive rights to AI output that another user could independently generate. This is a practical consequence of the lack of copyright protection, but it is also a contractual reality.
Some AI platforms retain a license to use your prompts and outputs for training purposes. If you are inputting confidential business information, trade secrets, or proprietary data into an AI tool, the platform's terms of service may grant it the right to use that data in ways you did not intend. Always review the AI tool's terms before using it for sensitive business work.
If your company builds AI-powered products that generate content for your customers, your own terms and conditions need to address output ownership. Who owns the content your AI generates for users? Can users claim copyright in outputs they prompt from your system? What warranties, if any, do you make about the originality or non-infringement of AI-generated output? These questions must be answered in your terms, not left to ambiguity.
Trade Secrets and Confidential Information: A Different Protection Path
Copyright is not the only form of IP protection. Trade secret law operates independently and does not require human authorship. If your business develops proprietary AI workflows, custom prompts, training data sets, or AI-generated analyses that provide a competitive advantage, trade secret protection may apply.
Trade secret protection requires that the information has independent economic value because it is not generally known, and that you take reasonable measures to keep it secret. If you develop a proprietary prompt engineering methodology that produces consistently better results than your competitors, that methodology can be protected as a trade secret even though the individual outputs it produces are not copyrightable.
To preserve trade secret protection, limit access to your proprietary AI workflows. Use NDAs with employees and contractors who work with these systems. Implement technical controls to prevent unauthorized access. And do not disclose your methodologies publicly unless you are prepared to lose trade secret protection over them.
What Happens Next: The Regulatory Landscape Is Still Moving
While the Supreme Court's refusal to hear Thaler settles the AI authorship question for now, Congress, the Copyright Office, and international bodies are all actively considering new frameworks for AI-generated works.
Several bills have been introduced in Congress that would create a new category of protection for AI-generated works, potentially with shorter terms or more limited rights than traditional copyright. None have passed as of mid-2026, but the legislative interest is real.
The Copyright Office has also initiated a multi-part study on AI and copyright, examining not just authorship but also the use of copyrighted works to train AI models, which is a separate and equally significant issue for businesses that develop or deploy AI systems.
For businesses, the practical takeaway is that the rules will keep changing. The policies and contracts you put in place today need to be flexible enough to adapt as the law evolves. Building your AI content strategy on the assumption that the current rules are permanent would be a mistake.
Frequently Asked Questions
If I use ChatGPT to write a blog post and publish it, do I own the copyright?
If you publish the raw AI output without significant human editing, no. That text has no copyright protection. Anyone can copy it. If you use the AI output as a starting draft and substantially rewrite, restructure, and add original content, the human-authored portions may be copyrightable. The extent of protection depends on how much original human creative expression is in the final version.
Can a competitor legally copy my AI-generated marketing materials?
If the materials were generated entirely by AI without significant human creative input, yes. Purely AI-generated content has no copyright protection and effectively exists in the public domain. This is why businesses that rely heavily on AI-generated content should implement human editorial processes to add copyrightable elements to the final output.
Does the AI tool I use affect my ownership rights?
Yes, in two ways. First, the AI tool's terms of service may impose specific conditions on your use and ownership of outputs. Some platforms retain rights to use your prompts and outputs for training. Second, the terms may disclaim any warranty of originality, meaning the same output could be generated for another user. Always read the terms of any AI tool before using it for commercial content creation.
How do I protect proprietary content if copyright does not apply to AI-generated portions?
Use a combination of strategies. Structure your workflow so that humans add significant creative expression to AI-generated drafts, making the final output eligible for partial copyright protection. Protect your proprietary AI workflows, prompt engineering methods, and training data as trade secrets. Use contracts (NDAs, employment agreements, vendor contracts) to control access and use of your AI-assisted work product.
What should my terms of service say about AI-generated content if I run an AI-powered platform?
Your terms should clearly state who owns AI-generated output (typically the user, but with important limitations). They should disclaim any warranty of copyright protection for purely AI-generated content. They should address whether identical output may be provided to other users. And they should limit your liability for any infringement claims related to AI output. A technology lawyer can draft these provisions to match your specific platform and risk profile.
Get Your AI Content Strategy on Solid Legal Ground
AI tools are transforming how businesses create content, code, and creative assets. But the legal framework has not caught up to the technology, and the default rules leave significant gaps that can cost your business real money and competitive advantage if you do not address them proactively.
Hansen Tong at TOS Lawyer advises businesses on intellectual property rights for AI-generated content, AI product terms of service, employment and contractor agreements for AI-assisted work, and IP protection strategies for companies that build or deploy AI tools. Contact TOS Lawyer to discuss how to protect your business's AI-related intellectual property and get your contracts and terms updated for the current legal landscape.
