Is writing the prompt enough to make your client the author?
It’s easy to see why prompt crafting could be confused with authoring a creative work—especially in today’s user-friendly artificial intelligence (AI) landscape. But according to Part 2 of the U.S. Copyright Office’s 2025 report on copyrightability, it’s not enough. A cleverly worded prompt, no matter how intricate or original, does not automatically result in a copyrightable output.
At the heart of the Copyright Office’s assessment is a simple idea: copyright protects human creativity, not machine outputs. Protectable authorship requires a meaningful, creative human contribution, not just instruction.
AI advancements are already reshaping how we make—and define—art, literature, design and even legal work product. Your clients’ expectations regarding output ownership, exclusivity and protection may need clarifying, and understanding where intellectual property rights begin and end is crucial for steering them away from costly misunderstandings.
This piece unpacks what the 2025 Copyright Office report says about AI and authorship—and why unpredictability matters. We’ll explore why prompts alone are insufficient for copyright claims, how courts are approaching AI-generated content and what legal professionals need to know when advising on copyrightability in an AI-driven world.
AI may be changing the deep things about us, like how we make art, but the legal system still demands a human hand on the pen, for now.
The U.S. Copyright Office’s current answer is clear: copyright law demands meaningful, creative human authorship— the act of directing a machine isn’t enough.
In Part 2 of its 2025 report on Copyright and Artificial Intelligence, the Office reaffirms a foundational principle: copyright protects human creativity, not machine outputs. The Copyright Clause in the Constitution speaks of securing rights to “authors” and “writings”—terms that have always been understood to require human expression. Courts have consistently reinforced this point, holding that the author of a copyrighted work must be the person who translates an idea into a tangible form.
So far, no U.S. court has recognized copyright protection for material generated entirely by non-human actors. For a work to qualify for copyright protection, there must be human creativity involved—and it must go beyond mere time, effort, or instruction. While noting that each determination will be unique, the Office outlines three broad scenarios where AI-assisted works may still be protectable:
These exceptions reflect a consistent theme: what matters is the human’s creative contribution, not the AI’s operation.
The Office directly addresses the growing misconception that prompting alone is enough of a creative contribution to trigger copyrightability. While writing a sophisticated or artistic prompt might feel creative, the generative process is inherently unpredictable. Identical prompts can produce wildly different results—and the internal workings of AI systems remain largely opaque. Without reliable, creative control over the output, it becomes difficult for the human user to claim authorship under traditional copyright standards.
The United States is not alone in wrestling with the copyrightability of AI-generated content. Around the world, countries are converging around the same baseline principle: copyright law exists to protect human creativity. But what qualifies as “human enough” remains context dependent.
Across the EU, there’s strong agreement that copyright protection only extends to AI-assisted works when the human input is significant. In 2024, the Council of the European Union surveyed member states and found widespread consensus that existing copyright laws were sufficient for addressing AI-generated outputs—so long as a human played a central role in the creative process. Fully machine-generated works? Not protectable.
Japan takes a fact-specific, case-by-case approach. According to 2024 guidance from its Copyright Subdivision of the Cultural Council, copyrightability of an AI-generated work depends on several factors: the depth of the user’s input, the number of generation attempts, the creative selection among multiple outputs and any human modification or enhancement of the result. Human involvement must be more than minimal.
Similarly, South Korea’s 2023 guidance asserts that “only a natural person can become an author.” Works created entirely by AI aren’t eligible for registration unless a human has made creative changes or re-arranged the material in a sufficiently original way. Notably, South Korea granted copyright registration to a film produced with AI assistance—but only because it had been substantively edited by a human author.
In a notable 2023 case, the Beijing Internet Court ruled that an AI-generated image was copyrightable under Chinese law. The court credited the human creator’s extensive role: issuing over 150 prompts, making detailed modifications and selecting from various outputs. While this ruling isn’t binding precedent, it signals a willingness to recognize human authorship where AI is clearly used as a tool, not the sole creator.
The UK occupies a unique position. Under an older statute still in force, works generated “by computer in circumstances such that there is no human author” may still be protected—with the person who arranged for their creation deemed the author. But whether this provision should apply to generative AI is currently under review. In 2024, the UK launched a new consultation to assess whether prompts or AI co-authorship should trigger protection.
Countries like Canada, India, Australia, New Zealand and Hong Kong also allow for protection of computer-generated works under certain conditions—but most are actively reviewing whether their frameworks are equipped to handle the complexities of generative AI.
Across these nations, public consultations, withdrawn registrations, and pending legislation all point to a shifting global landscape–one in which legal professionals must stay attuned.
While the U.S. Copyright Office ultimately determined that no new legal protections are currently needed for AI-generated works, its report hints that this may not always be the case. There are deeper tensions, and a series of policy debates that are quietly gathering force. The Office acknowledged several growing arguments for legal change.
Some commenters argued that granting copyright protection to AI-generated outputs would encourage innovation and cultural growth. However, most agreed with the Office that the Copyright Clause exists to incentivize human creativity—not machine production. They warned that extending protection to AI-generated material could ultimately disincentivize human authors and harm the creative ecosystem.
Others stressed that AI can serve as an essential tool for people with disabilities, enabling new forms of artistic expression. The Office strongly supported this goal—but emphasized that existing law already protects works where AI is used as a tool to assist human creativity, without the need for additional legal changes.
A few commenters raised alarms that a lack of protection for AI-generated content could disadvantage U.S. innovation on the global stage. The Office was unconvinced, responding that copyright should not be reshaped based on what other countries might or might not do—and that America's commitment to human-centered authorship remains a constitutional imperative.
Some stakeholders pushed for legislative reform, arguing that clearer statutory standards would help creators, businesses and courts navigate the rise of AI. The Office acknowledged these concerns but ultimately determined that copyright law’s flexible, case-by-case approach remains preferable for now.
Whether future changes will eventually lead Congress to create a new, tailor-made system—or a sui generis protection—for AI-generated works remains to be seen. For now, the U.S. legal system draws a bright line: authorship belongs to people.
As tempting as it may be to treat a prompt as a brushstroke, it isn’t currently enough to secure copyright protection. Clients experimenting with generative AI may assume they own their outputs, or that a clever prompt is enough to guarantee protection. It falls to counselors to explain where those assumptions break down—and to guide clients in building workflows that truly preserve their intellectual property rights.
In the co-creation age, attorneys must understand (and be able to explain) the difference between being creative and directing creation. Copyright was designed to protect and nurture the sparks of human ingenuity that eventually glow toward progress and prosperity. Generative AI may contain such sparks, perhaps worthy of fanning in their own right, but for now, we protect only the human hand that lights the flame.
