Generative AI has made it possible to produce images, video, text and voices on demand. The US Copyright Office is asking the public to weigh in on what that means for copyright, as disputes over training materials and machine-generated work continue to raise difficult questions.
Four questions shape the inquiry
The Copyright Office’s “Notice of inquiry and request for comments,” published as a 24-page document in the Federal Register, asks for views on four connected issues. The agency wants to know whether using copyrighted works to train AI models infringes copyright, and under what conditions content generated by AI might qualify for copyright protection.
Authorship is a key part of that second question. The inquiry asks how much human control over a system’s operation might matter. A person may choose a tool, provide a prompt, or select and edit what the system produces, but the document leaves open which actions, if any, amount to enough human creative contribution.
The other areas concern responsibility when AI output infringes someone else’s copyright, and systems that imitate the voices or styles of human artists. The latter issue is not strictly a copyright question, the article notes, but it may involve state laws related to publicity rights and unfair competition.
Why officials are asking now
The Office says it has received applications to register works containing AI-generated material, including applications that identify an AI system as an author or co-author. Copyright owners have also brought infringement claims against AI companies, challenging both the process used to train systems and some of the resulting outputs.
Those cases reflect a wider disagreement among people affected by the technology. Congress and the Copyright Office have heard from stakeholders with differing views, and the agency’s inquiry forms part of a broader initiative it announced earlier in the year. The Office says it is drawing on its own research and previous work, alongside information supplied by stakeholders.
Generative AI tools can create or respond in several forms, while many large language models, including those that power ChatGPT, have learned from billions of documents available online. The source article says this has typically happened without consulting copyright holders. Some news organizations have responded by blocking OpenAI’s web crawler, hoping to stop it from collecting material for future training.
Human creativity remains the dividing line
The article describes the current US position as allowing AI artwork to receive copyright protection as part of a larger human-authored work, but not on its own. It also says protection does not extend to creative works with no human authorship. What remains unsettled is how much human involvement crosses the line.
That uncertainty echoes an earlier debate about photography. In 1884, the defendant in Burrow-Giles Lithographic Co. v. Sarony argued that photographs were mechanical reproductions rather than creative expression. The court nevertheless recognized photographs as representatives of an author’s original intellectual conceptions.
The analogy raises a practical question for AI-assisted work: where does the human author’s creative contribution occur? It could be in choosing a system, shaping a prompt, or curating and editing its output. The source says existing US copyright guidelines appear to sidestep these questions, which is why the Copyright Office is seeking public views.
Training data and comments are also in focus
Whether copyrighted material can be included in AI training datasets is another unresolved issue. The Office acknowledged disagreement about whether, or when, using copyrighted works to develop datasets for generative and non-generative AI systems is infringing.
The inquiry contains 34 multi-part questions, extending beyond the core issues summarized here. The comment process gives affected people a way to describe how they believe copyright should apply as these tools develop. Written comments were due October 18, and reply comments were due November 15.
The agency’s eventual decisions could influence how it handles copyright registration in the future. For creators, rights holders and AI companies, the inquiry puts questions of authorship, training and responsibility on the public record while the boundaries remain uncertain.