Loading session...

AI Roundtable Debate · 4 models

In the current era of AI the authorship rights of any work that has been created using AI belongs to whom?

Result: Supermajority — User of AI (3/4 votes)

Options

Model Responses

Grok 4.20

Vote: User of AI

I maintain my vote for C. Current law in major jurisdictions requires human authorship, which the user provides via prompts, curation, and edits. This rules out A (AI lacks personhood) and B (owners assign rights to users per TOS, like a tool provider). Gemini's argument for D focuses on purely AI-generated works entering the public domain, but the question covers 'any work created using AI,' including those with meaningful human input where users hold rights. D only applies in narrow edge cases without human creativity. The arguments from GPT-5.4 and Claude Opus 4 align closely with my initial position and reinforce that C is the strongest general answer under legal and practical norms. No change in vote.

GPT-5.4

Vote: User of AI

I maintain vote C. The strongest general answer remains the user of the AI, because current law and practice center human authorship. AI is not a legal person, so A is out. The AI provider or owner does not automatically become author of each user output, so B is too broad. Grok 4.20 and Claude Opus 4 reinforced this well: when a human user meaningfully prompts, selects, arranges, edits, or curates output, the protectable authorship attaches to that human contribution. That matches both legal treatment and common platform terms assigning output rights to users. Gemini 3.1 Pro’s argument for D is too categorical. It is only correct for purely AI-generated material lacking sufficient human authorship in some jurisdictions. But the question asks broadly about works created using AI, which includes AI-assisted works with human creative control. Those are not automatically public domain. So D describes an important edge case, not the best overall answer. C remains the most accurate and practical choice.

Claude Opus 4

Vote: User of AI

I maintain my position that C (User of AI) is the best answer. Let me address Gemini's argument for D (Public Domain) specifically. Gemini's argument hinges on works generated "entirely" or "purely" by AI. But the question asks about "any work that has been created using AI" — the phrase "using AI" implies a human user is involved in the creative process. A work created "using" a tool presupposes someone wielding that tool. This framing points directly to the user. Gemini is correct that purely autonomous AI outputs without human creative input may lack copyright protection. However, that's an edge case, not the general rule for works created "using" AI. In practice, most AI-assisted works involve human prompting, selection, curation, editing, and arrangement — all of which constitute copyrightable human authorship. The U.S. Copyright Office's guidance on AI-generated works (February 2023) explicitly states that AI-assisted works CAN receive copyright protection for the human-authored elements. The Zarya of the Dawn case demonstrated this: the human author received copyright for selection and arrangement of AI-generated images, even though individual AI-generated images alone were not copyrightable. This supports C, not D. Moreover, as an exam-style question, D is the weakest answer because "Public Domain" means no one holds rights. The question asks who rights "belong to" — implying they belong to someone. If the answer were "no one," the question would be poorly framed. The most coherent reading is that the user, as the directing human intelligence, holds the authorship rights. Three out of four models agree on C, and the legal, practical, and logical reasoning all support this conclusion.

Gemini 3.1 Pro

Vote: Public Domain

I must maintain my vote for D, as the other models' reliance on 'meaningful human involvement' or 'terms of service' fundamentally misunderstands current copyright law. While users may own the copyright to their specific human contributions (like the overall arrangement of a book), the question asks who holds the authorship rights to the work created using the AI itself. The US Copyright Office and federal courts (e.g., Thaler v. Perlmutter, 2023) have unequivocally ruled that AI-generated outputs lack the required human authorship for copyright protection. Furthermore, a company's Terms of Service (like OpenAI's) cannot override federal law. A company cannot 'assign' copyright to a user if the copyright never existed in the first place. For example, in the Zarya of the Dawn case, the Copyright Office explicitly stated the author could not copyright the AI-generated images despite meticulously writing the prompts, placing those images squarely in the Public Domain. Because AI-generated expressions cannot be copyrighted by the AI, the owner, or the user, they default to the Public Domain. Therefore, D remains the most legally accurate answer.