Who Owns AI-Generated Content? The Copyright Question Still Unresolved
- Mehul Bansal
- 1 hour ago
- 3 min read
Adv Mehul Bansal, Jadetimes Staff
Who Owns AI-Generated Content? The Copyright Question Still Unresolved
Millions of people generate images, text, music, and video with AI tools every single day. Almost none of them know whether they actually own what they made, whether they can sell it, or whether they might be infringing someone else's rights in the process. That uncertainty isn't a gap in public knowledge it's a genuine, unresolved gap in the law itself.
The One Settled Rule: No Human, No Copyright
Copyright law has one firm answer so far: a work created entirely by AI, with no human creative input at all, cannot be copyrighted in the United States.
That rule was cemented by Thaler v. Perlmutter. Computer scientist Stephen Thaler tried to register a copyright for an image generated autonomously by his own AI system, listing the AI itself as the "author." The Copyright Office refused, a federal district court agreed, and in March 2025 the D.C. Circuit Court of Appeals affirmed that decision reasoning that the Copyright Act's own language, including references to an author's children, widow, and heirs, only makes sense if "author" means a human being. On March 2, 2026, the Supreme Court declined to hear a further appeal, leaving that rule firmly in place. One important nuance: declining to hear a case isn't a ruling on the merits it simply leaves the lower court's decision standing.
The Harder Question: What Counts as "AI-Assisted"?
Thaler was, by design, the easiest possible case: an AI system with zero human creative involvement. Most real-world AI use doesn't look like that. Someone writes a detailed prompt, generates dozens of variations, picks a favorite, and edits it further. Does that count as human authorship?
The Copyright Office's position is that prompting alone isn't enough even a highly detailed prompt is treated more like an instruction handed to a contractor than a creative act in itself. What matters is whether a human exercised genuine creative control over the work's expressive elements: composition, specific creative choices, and meaningful editing or arrangement after the fact. Purely AI-generated portions must be disclosed and excluded when registering a copyright; only the human-authored contributions can actually be protected.
That line is still being tested in court. Artist Jason Allen is challenging the Copyright Office's refusal to register "Théâtre D'Opéra Spatial," the Midjourney-generated image that won a Colorado State Fair art prize, arguing that his 600-plus iterative prompts, careful selection of style, color, and composition, and subsequent editing amounted to real authorship rather than just typing instructions. Both sides finished filing their summary judgment arguments in early 2026, and a ruling from the Colorado federal court is still pending. Whichever way it goes, it's likely to be the first real judicial test of exactly how much human involvement is enough.
A Separate Fight: What AI Was Trained On
Ownership of AI output is only half of the legal picture. A parallel, equally unsettled battle concerns whether AI companies were allowed to train their models on copyrighted material in the first place. Courts have not reached a consistent answer: one ruling found that training on lawfully acquired books was a transformative fair use, while relying on pirated copies of those same books was not. Other pending suits, brought by news organizations and publishers against AI companies over training data, remain unresolved, and at least one court outside the U.S. has ruled against an AI company on similar grounds. None of this changes the output-ownership rules above, but it means a business relying heavily on AI-generated content is currently navigating legal uncertainty from two directions at once, not one.
What This Means If You Use AI Tools
Editing, arranging, and refining AI output yourself creates a real, if partial, copyright claim simply downloading the first result generally does not.
Keep records. Your prompts, iterations, and the specific edits you made can matter later if you ever need to demonstrate your own creative contribution.
Copyright and commercial-use rights are separate questions. Even where your content isn't copyrightable, a platform's terms of service may still restrict or permit how you're allowed to use it commercially.
Assume anything purely AI-generated is legally unprotected, and could, in principle, be copied by anyone else.
The Bottom Line
For now, the safest working assumption is that AI is a tool you direct, not a co-author and the more visible your own creative decisions are in the final result, the stronger your position, both legally and practically.
Comments