Who Owns an AI-Generated Video? The 2026 Answer Is Uncomfortable
US law settled in 2026: purely AI-generated output has no human author and no copyright. What that means for creators in practice, and the one thing that changes it. Not legal advice.
The honest answer surprises most people, and it is now reasonably settled in the US.
This is general information, not legal advice. Rules vary by country, this area is moving, and if real money depends on the answer, ask a qualified lawyer.
The legal position
On 2 March 2026 the US Supreme Court declined to hear Thaler v. Perlmutter, which finalised the rule that US copyright requires a human author.
The consequence: if you produce a work solely from a prompt — however long and detailed that prompt is — the current consensus is that you do not hold copyright in the output. Purely AI-generated images and video are treated as being outside copyright protection in the US and much of Europe.
Which means, bluntly: a competitor can legally reuse your purely AI-generated clip.
The distinction that matters: ToS ≠ copyright
These are two different systems and people conflate them constantly.
Platform terms of service decide whether you may use the output commercially. A vendor can grant you broad commercial rights, and most paid tiers do — while free tiers frequently restrict output to personal, non-commercial use. That is a contract between you and the vendor.
Copyright decides whether you can stop other people using it. A platform saying "you own the output" cannot manufacture a copyright that does not legally exist.
So you can have full commercial rights to use a clip, and no ability to prevent anyone else using the same clip.
What restores protection: human authorship
The thing courts look for is human creative decision-making. Selecting among outputs, editing them, arranging them, rejecting the ones that miss, cutting several shots into a sequence, writing the script, directing the performance — these are authorship, and they are exactly what a real production process involves.
A single prompt with a single output is the weakest case. A piece where you wrote the script, chose the references, generated six shots, discarded four and edited the rest into a sequence is a much stronger one, and the protectable work is your selection and arrangement.
What creators should actually do
- Check the licence tier before commercial use. Free tiers commonly forbid it. Paid tiers commonly allow it. Read the specific plan.
- Do not treat a generated clip as a defensible asset. If a piece of footage is load-bearing for a brand, film it or commission it.
- Keep your human contribution visible and documented — script, edit decisions, shot selection. That is where any protection lives.
- Remember it cuts both ways. Other people's purely generated output is in the same position.
Where we sit
We run per-second endpoints and pass the output to you. Every model in our catalogue is billed by the second with no separate commercial tier, so the licence question does not turn on your plan.
The copyright question is not ours to solve and nobody in this category can solve it — it is a matter of law, not of terms. Anyone telling you their output is fully copyrightable because their terms say "you own it" is describing a contract, not a legal right.
The practical summary
You can sell it. You probably cannot stop others copying it. The way to make a piece defensible is the same as it always was: put enough of yourself into it that the work is recognisably authored — which is also what makes anyone watch.