Why Purely AI-Generated Works Cannot Be Copyrighted

Courts and copyright offices worldwide, most notably in the United States, have consistently ruled that works created entirely by artificial intelligence without human intervention cannot be copyrighted. This article explores the legal and philosophical reasons behind these decisions, focusing on the fundamental requirement of human authorship, the constitutional purpose of copyright law to incentivize human creativity, the lack of traditional creative control in prompt engineering, and the practical legal dilemmas of assigning ownership to non-human entities.

The Fundamental Requirement of Human Authorship

The primary reason AI-generated works are denied copyright protection is that copyright law is historically and legally rooted in human creation. In the United States, the Copyright Act protects "original works of authorship." While the statute does not explicitly define "author," courts and the U.S. Copyright Office (USCO) have long interpreted this term to mean a human being.

Historical legal precedents support this interpretation. Most famously, in the "Monkey Selfie" case (Naruto v. Slater), a federal appeals court ruled that a photograph taken by a macaque monkey could not be copyrighted because animals do not have legal standing under the Copyright Act. Courts apply this same logic to artificial intelligence. In the landmark 2023 case Thaler v. Perlmutter, a federal judge upheld the USCO's refusal to register a visual artwork created autonomously by an AI system, stating unequivocally that "human authorship is a bedrock requirement of copyright."

From a philosophical and economic standpoint, copyright law exists to promote the progress of science and useful arts by granting creators a temporary monopoly over their works. This financial and legal incentive encourages humans to invest time, effort, and resources into creating new art, literature, and music.

Machines, algorithms, and AI models do not respond to economic incentives. An AI does not need the promise of royalties or legal protection to generate an image or write a piece of code; it simply executes its programming when prompted. Extending copyright protection to purely machine-generated outputs does not fulfill the constitutional purpose of inspiring creation, as the machine has no consciousness or motivation to be incentivized.

Lack of Creative Control in Prompt Engineering

Many users argue that entering text prompts into an AI system is akin to a photographer using a camera, a precedent established in the 1884 Supreme Court case Burrow-Giles Lithographic Co. v. Sarony, which granted copyright to photographs. However, copyright offices make a sharp distinction between using a tool and delegating the creative process to an algorithm.

When a photographer takes a picture, they retain strict control over lighting, angles, shutter speed, and composition. Conversely, when a user inputs a text prompt into a generative AI model, the AI makes the ultimate "creative" decisions regarding how to interpret that prompt. The USCO has stated that because the user cannot predict or control the exact output of the AI, supplying a prompt is more like commissioning a human artist than creating the work oneself. Without direct human control over the expressive elements of the final product, authorship cannot be established.

Granting copyright to autonomous AI works would also create severe logistical and legal problems within the intellectual property system.

  • Ownership Ambiguity: If an AI work were copyrightable, the law would struggle to determine who owns the rights. It is unclear whether the copyright would belong to the person who wrote the prompt, the company that coded the AI, or the original creators of the vast datasets used to train the model.
  • Copyright Duration: Copyright terms are frequently tied to the lifespan of the author (in the U.S., the life of the author plus 70 years). Because an AI does not have a biological lifespan, calculating the expiration of the copyright under current legal frameworks is impossible without legislative intervention.
  • Market Flooding: AI can generate thousands of texts, images, and musical compositions in seconds. Allowing these works to be copyrighted would allow individuals or corporations to monopolize vast swaths of creative expressions instantly, potentially stifling human creators who might accidentally infringe upon millions of AI-generated copyrights.

While works that combine human creativity with AI assistance may receive partial protection for the strictly human-made elements, purely AI-generated works remain firmly in the public domain.