It depends on what you mean by “creating.” A person may choose the goal, give directions, select an output and revise it, while an AI system generates some or all of the expressive material. For copyright, the key question is narrower: what expressive choices did a human make, and what does the law in the relevant jurisdiction protect?
“Creating” can mean several different things
In everyday use, the person who asks an AI system to make an image or text may reasonably say they created it. They initiated the task and may have shaped the result. But that does not settle who determined the work’s expressive details, who qualifies as its author under copyright law, who owns any rights, or who is responsible for publishing it.
As an Amazon Associate I earn from qualifying purchases.
Keep those questions separate:
- Direction: Who chose the goal, subject or constraints?
- Expressive control: Who determined the specific words, forms, composition or other expressive elements?
- Legal authorship and protection: What does copyright law recognize in that jurisdiction and for that kind of work?
- Publication and disclosure: Who publishes or deploys the material, and do transparency rules apply?
These roles can belong to different people or entities. Calling someone the “creator” in a social or product context does not, by itself, answer the legal questions.
Free tools Windows power users keep installed
One-click scans. No signup required.
Does writing a prompt make you the author?
In the United States, a prompt alone is not enough
The U.S. Copyright Office’s January 29, 2025 report on copyrightability of generative AI outputs says protection depends on whether a human determined sufficient expressive elements of the resulting work. The Office says merely providing prompts does not, on its own, establish that level of human authorship. Asking for a subject, style or outcome may guide a system without determining the particular expression it produces.
#1 Best Overall
That does not mean a person who uses AI can never hold copyright in a work. The Office identifies human-authored material that is perceptible in the result, a sufficiently creative arrangement of human and AI-generated material, and creative human modifications as possible grounds for protecting the human contribution. The focus is on what the person actually contributed, rather than on whether AI appeared anywhere in the process.
The UK has a different rule for certain computer-generated works
The UK government’s 2026 Report on Copyright and Artificial Intelligence describes section 9(3) of the Copyright, Designs and Patents Act 1988. For a qualifying computer-generated literary, dramatic, musical or artistic work with no human author, the person who made the arrangements necessary for its creation is treated as the author. The report says that, in a general-purpose AI prompt scenario, this will usually be the person who input the prompt. It describes the term of protection for such a work as 50 years from its creation.
This is a jurisdiction-specific statutory rule, not a general principle that a prompt makes someone the author everywhere. Nor should it be confused with the U.S. Copyright Office’s test for whether human-determined expression is present.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesThe EU approach described in the UK report centers on human originality
The 2026 UK report describes EU copyright originality as requiring an author’s intellectual creation expressed through free and creative choices. It says EU law does not provide the same specific protection for computer-generated works without a human author as the UK rule. This is a comparative orientation, not a determination of how a particular work or claim will be treated in every EU member state.
How the approaches compare
| Jurisdiction | Human creative control | Work with no human author | What to keep in mind |
|---|---|---|---|
| United States | The U.S. Copyright Office’s January 2025 position requires sufficient human-determined expressive elements. Human-authored material, creative arrangement or creative modification may qualify; prompting alone does not. | The materials cited here do not establish a special rule equivalent to the UK computer-generated-works provision. | The Office says existing law is adequate and additional protection for AI-generated output is not warranted on the record it reviewed. This is the Office’s stated position, not a universal rule for other jurisdictions. |
| United Kingdom | For works with human authorship, the report discusses human contribution; for qualifying computer-generated works with no human author, section 9(3) assigns authorship to the person making the necessary arrangements. | Yes, for qualifying literary, dramatic, musical or artistic works under the statutory rule described in the 2026 report. | The report describes the term for a computer-generated work as 50 years from creation. It also discusses possible reform; those options are not enacted changes. |
| European Union | The UK report describes the originality standard as requiring an author’s intellectual creation and free and creative choices. | The UK report says EU copyright does not provide specific protection for computer-generated works without a human author. | AI Act transparency duties are a separate issue from copyright authorship and ownership; applicable duties depend on the actor, content and circumstances. |
The comparison reflects the U.S. agency materials and the UK government’s 2026 account of UK and EU law. Other countries may take different approaches, and individual outcomes can depend on the work, the facts and the relevant law.
AI assistance does not automatically erase human authorship
An AI-assisted work can include protectable human expression even if a system generated some of its material. For example, a person may contribute original prose and use AI to help with a passage, or make creative edits to generated material and combine it with human-authored work. Under the U.S. Copyright Office’s 2025 explanation, the relevant question is whether the human-authored expression, arrangement or modification meets the applicable standard—not whether the workflow was entirely AI-free.
That principle also helps avoid the opposite mistake: treating every person who operates an AI tool as the author of everything it produces. A human’s involvement matters, but its nature and extent matter too.
Copyright authorship is not the same as disclosure or responsibility
Whether a person qualifies as an author is distinct from whether they own rights, whether a platform or employer has contractual claims, and who bears responsibility for what is published. The materials summarized here do not resolve ownership or liability for every arrangement; those questions can involve other law and the terms governing a tool or project.
Transparency rules answer a different question: whether generated or manipulated content must be marked or disclosed. The UK government’s 2026 report describes EU AI Act requirements that include machine-readable marking for certain synthetic outputs and disclosure by deployers in defined deepfake and public-interest contexts, subject to exceptions. A label does not, by itself, establish who made the expressive choices or who owns copyright. For a real compliance decision, check the applicable EU rules and implementation guidance for the specific actor and use.
What current policy discussions do—and do not—show
The U.S. Copyright Office says more than 10,000 responsive comments informed its conclusions about copyrightability. That number counts submissions, not people in a representative poll and not a measure of how much humans or machines create.
The UK government’s 2026 report says 78% of online survey respondents who answered the relevant consultation question opposed maintaining current computer-generated-works protection. Fewer than half of consultation respondents answered the relevant questions, so that figure describes a subset of participants, not the UK public as a whole. The report considers policy options; it does not establish that the UK rule has been changed.
Recommended Free Tools
Neither figure measures the philosophical or social question of who is “actually creating.” They concern, respectively, agency comments on copyrightability and responses to a policy consultation.
Where the AI-training debate fits
Who may claim copyright in an AI output is not the same issue as whether material used to train an AI system was lawfully used. The U.S. Copyright Office treated training and licensing separately from its Part 2 report on output copyrightability. Its initiative page said a pre-publication version of Part 3 was released on May 9, 2025, with a final version to follow. The output-authorship conclusions do not, by themselves, determine whether particular training uses require permission or infringe copyright.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
