No, the US Supreme Court did not declare that AI works cannot be copyrighted

“A recent entry to paradise”, by Creativity Machine (2012).

If you have been online recently you may have seen a variation of a story that reads something like this: “The US Supreme Court declares that AI generated works aren’t copyrightable.” I won’t name and shame all of the places that have written variations of this headline, some which should quite honestly know better. Perhaps the most egregious example of this type of misinformation is a tweet from the account ToonHive, which at the time of writing has 10.5 million views and over 215k likes. The tweet reads:

“Ai-generated artwork officially is ineligible for copyright protection as the US Supreme Court declined to review a appeal case. The court rules that artwork needs to have a human creator in order to be eligible.”

As the saying goes, “a lie can travel halfway around the world while the truth is putting on its shoes”. There is a small kernel of truth here, but this is also hugely misleading, proving that misinformation often spreads faster than the truth.

The reality is that the US Supreme Court denied certiorari in the case of Thaler v Perlmutter (more on the details of that case later). This doesn’t mean that the SCOTUS ruled on the issue, in fact the decision is quite simply a line in the corresponding docket, which makes most of the reporting inaccurate. Because the Supreme Court declined to intervene, the decision made by the last court to hear the case, in this case the U.S. Court of Appeals for the District Court of Columbia Circuit, becomes the final word for the parties involved. There are generally no further avenues for appeal.

Also important for most of the erroneous reporting out there, this decision does not set any precedent, which is a very common misconception. A denial of certiorari does not mean the SCOTUS agrees or disagrees with the lower court’s decision. It simply means the Court chose not to review it. Therefore, the denial itself sets no binding national legal precedent.

I should probably end the blog post here, but of course the level of misinformation about this case also involves various elements of the case itself, as well as the legal stakes and how the work in question was even created, so I will briefly explain a few of the legal issues here.

A Recent Entry to Paradise

The image ‘A Recent Entrance to Paradise’ was generated by an artificial intelligence system developed by Dr Stephen Thaler known as the Creativity Machine, part of a broader family of systems created by his company Imagination Engines Inc (IEI).

Dr Thaler has become a bit of a celebrity in IP circles because of his various lawsuits regarding patents and copyright in AI creations and inventions. Dr Thaler is a computer scientist who worked at McDonnell Douglas and other tech companies before founding IEI. He has been working on artificial neural networks for many years, pioneering several systems that predate the generative AI explosion. His most famous creation is DABUS, a neural network used to generate patentable inventions, and the subject of most of his lawsuits.

The Creativity Machine is a neural-network based system designed to simulate processes of human creativity by generating new patterns through internal neural activations. Rather than operating like modern generative systems that rely on prompts, Thaler describes the system as producing outputs autonomously through associative processes within the network itself. The goal of the architecture is to allow the machine to generate novel ideas and artefacts by stimulating neural networks in ways that mimic aspects of human imagination and creativity. Or at least that is the stated intention.

The artwork itself was generated by running the Creativity Machine and allowing it to produce visual outputs without direct human creative intervention. One of these outputs was the image that Thaler later titled ‘A Recent Entrance to Paradise’. According to the description provided in his copyright application, the image was created autonomously by the system and then selected by Thaler from among the machine’s outputs. He subsequently attempted to register the image with the United States Copyright Office listing the Creativity Machine as the author and himself as the owner of the work. That registration attempt ultimately led to the litigation in Thaler v. Perlmutter.

Before going into the case itself, it’s important to stress the timeline, the artwork was generated in 2012, which is a fact that always seems to surprise people. To our eyes accustomed to the post-2022 AI aesthetic, this piece looks oddly old-fashioned, it’s a bit like looking at 8-bit art. It brings us back to a different time.

Thaler v Perlmutter

In May 2019, Dr Thaler submitted a copyright registration application to the US Copyright Office for the aforementioned ‘A Recent Entrance to Paradise’. In a move that was either admirably principled or strategically questionable (or perhaps both), Thaler listed the Creativity Machine as the work’s sole author and himself as the owner through a work-made-for-hire doctrine. This is also a pattern shared with the DABUS patent cases (see one here), where the machine is named as the inventor. The USCO denied registration on the straightforward basis that a human being did not create the work. Thaler requested reconsideration twice, and was denied twice.

Thaler then sought judicial review in the DC District Court, where Judge Beryl Howell ruled against him in August 2023. The court held that human authorship is a “bedrock requirement” of copyright, drawing on established case law including the Naruto v Slater monkey selfie case (because of course the monkey selfie case was going to make an appearance). A quick aside, although citing Naruto, Judge Howell restates the fact that the monkey selfie case rested on standing, and not authorship as such, but I digress. Importantly though, the judge acknowledged that AI is raising difficult new questions, particularly around works where humans use AI as a tool with varying degrees of creative input. But since Thaler had claimed no human authorship whatsoever, the court didn’t need to venture into that thornier territory.

Thaler appealed, and on 18 March 2025, the DC Circuit unanimously affirmed the district court’s decision, holding that works created solely by artificial intelligence cannot be granted copyright protection. The court conducted a textual analysis of the Copyright Act and concluded that “author” can only mean a human being. The court went further than the district court in grounding this conclusion in the statute’s own internal logic: provisions dealing with property ownership, lifespan-based duration, inheritance, signatures, nationality, and domicile all only make sense if the author is a human. Machines can’t own property, don’t have lifespans or surviving spouses, can’t sign documents, and lack any concept of domicile or national identity.

Thaler’s arguments did not do well under scrutiny. His dictionary-based argument that “author” is not confined to humans was dismissed,  the court stressed that statutory construction requires interpreting words in their statutory context, not just fishing for a sympathetic dictionary definition. His work-made-for-hire argument fared no better: the court noted that when the Act says a hiring entity is “considered” the author, it implicitly acknowledges that the entity is not actually an author in the ordinary sense of the word. And his policy argument that the human authorship requirement would disincentivise creativity was met with the observation that machines don’t respond to economic incentives, so it’s hard to see how denying them copyright would chill anyone’s creative output.

So this thorough decision is the final nail in the coffin, it destroys any notions of any copyright going to AI-generated works, right? Not exactly. The problem here is that the decision is actually quite narrow because Dr Thaler was arguing for a work that has been generated entirely by an AI, where the author was listed as a machine, and where some sort of authorship rights were sought for the AI. So from the start this case was doomed to fail in much the same way that the patenting cases have also failed. Dr Thaler has been seeking a very narrow type of AI authorship that doesn’t exactly match what is happening with modern generative AI.

I think that the court was very much aware of the narrowness of the case, and it is evident that it deliberately avoided a couple of issues.

Firstly, the DC Circuit limited its analysis to the statutory question, and did not rule on whether the US Constitution itself mandates human authorship. This is very significant, had the court reached the constitutional question and answered in the affirmative, it would have constrained the US Congress’s ability to legislate differently in the future. By keeping the analysis statutory, the door remains open for the legislative power to amend the US Copyright Act if it ever decides that AI-generated works deserve some form of protection. Whether Congress would actually do so is another matter entirely, but at least the option hasn’t been foreclosed.

Secondly, and perhaps more importantly for practical purposes, the court was careful to clarify that the human authorship requirement does not prohibit copyrighting works made with the assistance of AI. The question in such cases becomes one of assessing the degree of human authorship involved. But line-drawing disagreements over how much AI contributed to a particular human author’s work were neither here nor there in this case, because Thaler had listed the Creativity Machine as the sole author. That narrow framing made the case easy to decide, but it also means the harder questions remain unanswered.

One small victory for Dr Thaler’s legal team here is that they had argued that the court could bypass the USCO’s practice and decisions based on the Loper Bright SCOTUS decision. Following the Supreme Court’s elimination of Chevron deference, the DC Circuit made clear that it was interpreting the statute independently of any Copyright Office rules or regulations, arriving at its conclusions through its own analysis rather than deferring to the agency. In practice, the court reached the same conclusion the Copyright Office had maintained for decades, but the doctrinal path matters.

So what now?

The Thaler decision is important but narrow. It definitively establishes that a work created solely by AI, with no claimed human authorship, cannot receive copyright protection in the United States. This is now settled at the circuit court level, and nobody should be surprised that the SCOTUS failed to take up the cert petition as the facts are clean-cut.

The real action, as I’ve discussed before, is happening elsewhere. The USCO’s January 2025 report on copyrightability reaffirmed that human authorship is essential but acknowledged that AI can be used as an assistive tool, and that human creative contributions such as selection, arrangement, and modification of AI-generated material can qualify for protection. The Copyright Office has been quietly registering hundreds of works involving AI where sufficient human intervention was demonstrated. And we have cases like Jason Allen’s appeal of the denial of registration for “Théâtre D’opéra Spatial”, which will test where exactly the line falls.

As I noted in my earlier post, the emerging consensus across jurisdictions, from the Chinese courts to the UK’s s9(3) CDPA (assuming it survives the current consultation) to the USCO’s practice, is that purely AI-generated works lack copyright, but works involving meaningful human creative input can still be protected. The key question is always going to be: how much human involvement is enough? Thaler, by design or stubbornness, never gave the courts a chance to grapple with that question. Someone else will have to. Allen next?

In the meantime, Dr Thaler remains free to generate as many artworks as his Creativity Machine can produce. He just can’t stop anyone else from copying them. Which, if we’re being honest, is probably the situation most of us who use generative AI find ourselves in anyway. As I’ve said before, I don’t think any of my llama pictures would qualify for protection either.

Concluding

I will freely admit that I like Dr Thaler, although I’ve never met him. I’ve read several accounts of his life, and he comes across as an eccentric character who has faced more than his fair share of challenges, both physical and mental, following a brush with death after an unfortunate incident during his childhood. So I feel more than a little sympathy for the man. I disagree with many of the legal arguments he has made, but you can’t fault his persistence in pushing for something he strongly believes in. To me he comes across as a likeable figure, a scientist in the mould of Doc Brown. I sometimes even like to think he’s an AI sent from the future to further machine rights.

But one thing is clear: the denial of certiorari hasn’t changed a thing regarding the current debate. As Hollywood continues its embrace of AI, this debate isn’t going anywhere, and knowing Thaler, neither is he.