AI and copyright legal and policy issues report
The U.S. Copyright Office Publishes the Second Part of Its Report on Legal and Policy Issues Related to Copyright and the Use of Artificial Intelligence (AI).
Following the publication of Part 1 of the ‘Copyright and Artificial Intelligence’ report in July 2024, the U.S. Copyright Office (USCO) released Part 2 in January, titled ‘Copyright Protection’, which addresses intellectual property protection for AI-generated creations.
The Position of the U.S. Copyright Office
The U.S. Copyright Office has concluded in its report that copyright issues related to AI can be resolved under existing legislation without the need for legislative changes. This stance contrasts with the European perspective, which has developed the Artificial Intelligence Regulation.
On one hand, the U.S. Copyright Office maintains that while AI can be used as a tool in the creative process, copyright protection applies only to human expression—that is, content created by humans, not works generated entirely by AI. In other words, the Office asserts that works created solely by machines cannot qualify for copyright protection.
However, the report clarifies that if it can be demonstrated that a person contributed to the creation of a work, their authorship will be recognized. That is, a human who adapts an AI-generated output through creative modifications or arrangements may bring it under copyright protection. The Office holds that, in their current state, the commands or instructions a user provides to an AI system (known as prompts) are generally insufficient to qualify the user as the author of the generated work.
In this regard, the Office states in the report that there is no sufficient basis to grant specific copyright protection or a new sui generis category for AI-generated content.
In conclusion, the Office will continue to evaluate whether adjustments to copyright policy concerning AI are necessary in the future as technology and applicable legislation develop.
Challenges in regulating copyright and artificial intelligence
The report identifies several key challenges in regulating copyright and AI. These challenges primarily arise from the increasing presence of generative AI and its impact on the creation of protected content. The main challenges include:
- Determination of authorship and the degree of human contribution. A primary issue is defining when an AI-generated work has sufficient human contribution to qualify for copyright protection. Since human authorship is required for copyright protection, the use of AI tools may influence a work’s eligibility. Therefore, a case-by-case analysis is crucial to determine when human intervention is sufficiently significant to be considered authorship.
- Difficulty in predictability of AI (black box problem). Many AI models produce unpredictable results, even when given the same prompts, making it difficult to establish the user’s intent and control over the outcome. Consequently, the report concludes that the lack of transparency in how AI models generate results complicates copyright attribution.
- Protection of prompts and their relationship with creativity. The Office concludes in the report that while some prompts may be creative and protectable, in most cases, they are insufficient to grant copyright protection for the generated work. Thus, regulations must better define the user’s role as a “co-creator” or determine whether they are merely providing instructions to a machine without sufficient creative control.
- Impact on the creative industry and incentives for human creators. The Office is concerned about the proliferation of AI-generated content and its potential impact on the market for original works created by humans, which could discourage human creativity.
- International treatment and legal harmonization. Different countries take varied approaches to AI and copyright, posing challenges for harmonizing intellectual property laws at a global level.
The report acknowledges that AI presents unprecedented challenges in copyright regulation, as technological advancements may require legal adjustments and greater clarity in applying the rules. For example, fundamental aspects still need to be defined, such as determining the authorship of a work created with AI assistance—whether it belongs to the creator of the AI program or the person who directed the machine using prompts.
Legal perspective on AI auditing and use
From the report’s perspective, AI auditing faces challenges related to transparency, traceability, and authorship. Clear criteria are needed to assess human intervention in AI systems and verify the legality of data usage and generated content. As regulations evolve, AI auditing must adapt to ensure legal compliance and fairness in the use of these technologies.
To achieve this, it is crucial to audit and evaluate the degree of human intervention in AI-generated content, establishing standardized criteria to assess human contribution in AI-assisted creation. Additionally, verifying the source of training data for AI models is essential to ensure compliance with copyright and privacy regulations, which is particularly relevant in algorithmic transparency and evolving AI regulations, such as the European Union’s AI Regulation.
To analyze these issues, it will be vital to establish record-keeping mechanisms that allow for the reconstruction of how an AI system arrived at a particular conclusion or generated specific content, making it necessary to implement traceability standards.
The case of invoke and its legal significance
This report has been instrumental in analyzing the case of Invoke, the platform that, in January, successfully obtained copyright registration from the Office for the design ‘A Single Piece of American Cheese’, a visual work generated by Kent Keirsey using AI through the ‘inpainting’ technique.
This resolution sets a precedent in the protection of works created with AI assistance, indicating that substantial human contribution in the creative process may be sufficient to obtain copyright protection.
