Courts Settled the Easy AI Copyright Question. The Hard One Has No Answer Yet.
What happened
On March 18, 2025, the US Court of Appeals for the District of Columbia Circuit issued its ruling in Stephen Thaler v. Shira Perlmutter and put an official end to the narrowest version of the AI copyright debate: no, a machine cannot hold copyright, and a human cannot acquire copyright simply by owning a machine that produced a work. The principle that copyright requires human authorship held.
Thaler had been pressing this argument since 2018, when he applied to register a work he said was created entirely and autonomously by an AI system he built. The Copyright Office refused the registration on the grounds that copyright protection has always required a human author. Thaler sued, arguing that the statute did not mandate human authorship and that the Copyright Office was reading a requirement into law that Congress had never written. The district court sided with the government in 2023. The DC Circuit, after oral argument in September 2024, affirmed.
The court's reasoning was not that Congress had explicitly excluded machines but that human authorship is so foundational to copyright's history and purpose that it carries implicit force throughout the statute. Copyright, as the court understood it, was designed to encourage human creative expression by giving creators an economic incentive. A system with no stake in economic incentives, and no standing to hold property, falls outside the framework whether or not a drafting committee ever wrote that exclusion down.
The case Thaler pressed was always the clean version of a much harder problem. Sole autonomous AI authorship, with a named machine and a human owner who wants to claim downstream rights, is nearly absent from everyday practice. What is common, and growing more common, is partial AI generation: works where a person uses an AI tool to draft, extend, remix, or refine creative output and then releases the result under their own name. The Copyright Office and courts are fielding those questions now, with no settled doctrine and no consistent threshold for what counts as meaningful human contribution.
That uncertainty is where the documentation gap sits. When a creator claims copyright in an AI-assisted work, the degree of human contribution is the deciding fact, and right now nothing requires that fact to be captured at the moment the work is made. A provable record of what a system produced versus what a human changed, timestamped at creation rather than reconstructed during litigation, would make those disputes testable. Without it, every claim about how much a person actually contributed reduces to assertion, and courts are left deciding questions of authorship with no evidence that touches the actual creative act.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.