The legal fight over whether an AI system can be treated as an inventor has hit another wall. The UK Supreme Court has ruled against US computer scientist Stephen Thaler, rejecting his attempt to register two patents for inventions he says were created by his AI system, DABUS.
The UK decision
The Supreme Court ruled unanimously that AI cannot be considered an "inventor" for patent purposes. That conclusion upheld the position taken earlier by the UK's Intellectual Property Office (IPO), which rejected Thaler's application on the basis that an inventor must be a person or a company.
The dispute centered on two patent applications connected to DABUS. Thaler did not present himself as the inventor. Instead, he sought recognition for the AI system itself, making the case a direct test of whether patent systems can accept machine inventorship.
The court's answer was no. For now, the UK position remains that patents require an inventor who fits within existing legal categories, and an AI system does not meet that standard.
A pattern across jurisdictions
The UK ruling does not stand alone. The Supreme Court's decision is in line with similar rulings in Europe, Australia and the United States, where Thaler has also faced setbacks in efforts to secure legal recognition for AI-generated inventions.
Thaler's lawyers argue that the decision exposes a weakness in UK patent law. In their view, the law is not adequate to protect AI-generated inventions or to support industries that rely on AI for technology development.
That argument points to the larger issue beneath the case. AI systems are increasingly being used in creative and technical work, but legal frameworks still rely on older assumptions about people, authorship, inventorship and ownership. The Thaler cases force courts to answer a narrow question: can the machine itself be named as the legal source of the work?
So far, in the cases described here, courts have largely declined to take that step.
Copyright cases raise a related question
Thaler has also pursued copyright claims involving AI-generated content. In August, U.S. District Judge Beryl A. Howell rejected his attempt to enforce copyright on AI-generated content, siding with the U.S. Copyright Office's position that AI images cannot be copyrighted.
That case involved the AI image "A Recent Entrance to Paradise," which Thaler argued was a commissioned work by his AI system, the Creativity Machine. Thaler is a member of the Artificial Inventor Project.
Judge Howell's reasoning centered on human authorship. She argued that human authorship is a fundamental requirement for copyright. At the same time, she noted that current copyright law is reaching new frontiers with AI in the arts.
That matters because the ruling did not settle every copyright question involving AI. Instead, it left open a future issue: how much human involvement is required before an AI-generated work can be protected by copyright?
Why the Thaler cases have limits
The patent and copyright rulings are important, but they do not fully answer the broader AI copyright debate. A key reason is the way Thaler framed the cases. He deliberately insisted that he had no involvement in creating the image and that the machine was the inventor or creator.
That framing gave courts a relatively direct question to answer. If there is no human creator or inventor in the claim, can the AI system itself receive legal status? In the UK patent case and the US copyright case described here, the answer was no.
But many real-world AI disputes may be more complicated. Future cases may involve people who guide, select, revise or otherwise shape AI-generated works. Judge Howell noted that future rulings will need to clarify the level of human involvement required for AI-generated works to be copyrightable.
That is where the legal debate is likely to become more difficult. The question may shift from whether a machine alone can own or create protected work to whether a person's intellectual contribution is enough when an AI tool is part of the process.
A different result in Beijing
The source article also points to a contrasting decision from the Beijing Internet Court, which recently recognized the copyright of an AI-generated image.
In that case, the plaintiff had used Stable Diffusion to create a photorealistic image and shared it on a social platform. A blogger then distributed the image without permission, leading the plaintiff to file a copyright infringement lawsuit.
The court ruled that the AI-generated image was copyrighted because it reflected the plaintiff's original intellectual investment. The decision may serve as a precedent for future copyright disputes involving AI-generated content.
Taken together, these cases show why the AI copyright debate is far from settled. Courts are drawing lines around AI inventorship, AI authorship and human involvement, but those lines are not identical everywhere. For companies, creators and industries using AI in technology development, the central question remains practical as well as legal: when AI helps produce something valuable, who can claim protection, and on what basis?