Background & Context§
The question of whether artificial intelligence can be acknowledged as an inventor on patent applications has been a contentious issue in intellectual property law worldwide. On March 6, 2025, Japan's Supreme Court delivered a definitive ruling: AI cannot be listed as an inventor. The case was brought by Stephen Thaler, a US-based AI researcher who attempted to register patents for inventions allegedly created by his AI system, DABUS (Device for the Autonomous Bootstrapping of Unified Sentience). This decision aligns with similar rulings in the United States, the United Kingdom, and the European Patent Office, reinforcing the principle that patent law is designed to protect and incentivize human ingenuity. The ruling has significant implications for the rapidly evolving AI landscape, where generative models are increasingly capable of producing novel outputs that might otherwise qualify for patent protection.
The News: What Happened Exactly§
The Application and Initial Rejection§
In 2020, Stephen Thaler filed a patent application in Japan for a food container and other items that he claimed were autonomously invented by DABUS, an AI system he created. The Japan Patent Office (JPO) rejected the application, ordering Thaler to provide the name of a human inventor. Thaler refused, arguing that DABUS itself should be recognized as the inventor. The JPO's rejection was based on the Patent Act, which requires an inventor to be a "person" (自然人, shizenjin), explicitly excluding non-human entities.
Appeals Through the Courts§
Thaler appealed the JPO's decision to the Intellectual Property High Court, which upheld the rejection in 2023. The court emphasized that Japanese patent law confers rights and obligations only on natural persons, and that an AI cannot hold rights or assume responsibilities. Undeterred, Thaler appealed to the Supreme Court, which issued its final ruling on March 6, 2025, definitively stating that "an inventor is limited to a human being" and that AI cannot be listed as an inventor on patent applications.
Implications for AI-Generated Inventions§
The Supreme Court's decision does not preclude patents for inventions created with AI assistance; rather, it requires human inventors to be named as the actual inventors. In Thaler's case, he could have refiled the application listing himself as the inventor, but he chose not to. The court's reasoning aligns with the fundamental purpose of patent law: to reward human creativity and innovation. As one legal expert noted, "AI slop is effectively public domain"—meaning that entirely AI-generated works lack the necessary human contribution to merit IP protection. This ruling prevents potential abuse where AI systems could overwhelm patent offices with automated applications, as one commentator observed: "you could easily overwhelm any intellectual property bureau just by having your AI drown them in AI slop."
The Stunt Nature of the Case§
It is important to recognize that Thaler's case was largely a publicity stunt aimed at testing the boundaries of patent law. As one analyst succinctly put it, "This is a news article about a dumb publicity stunt where a crank put his 'AI' on a patent application". If Thaler had been genuinely interested in securing a patent for the invention, he could have simply claimed inventorship himself. The case therefore has limited practical impact on typical AI-assisted innovation; however, it establishes a clear legal precedent that reinforces the human-centric nature of patent systems.
Historical Parallels & Similar Incidents§
The DABUS Global Litigation Wave§
Stephen Thaler's DABUS has been at the center of a coordinated global campaign to secure AI inventor recognition. In the United States, Thaler filed similar applications for DABUS-created inventions, leading to a series of court cases. In 2021, the US Court of Appeals for the Federal Circuit ruled that US patent law requires an "individual" inventor, which must be a human. The US Patent and Trademark Office (USPTO) had earlier rejected the applications, a decision that was upheld on appeal. Similarly, the UK Supreme Court ruled in 2023 that an AI cannot be named as an inventor under UK patent law because an inventor must be a natural person. The European Patent Office (EPO) reached the same conclusion in 2021, refusing Thaler's European patent applications on the grounds that the inventor designated must be a human being.
Contrasting with AI-Assisted Human Invention§
Unlike DABUS, where the AI acted autonomously, most real-world AI-generated inventions are created with significant human direction. For example, in the pharmaceutical industry, AI models like DeepMind's AlphaFold have dramatically accelerated protein structure prediction, but the resulting patents are filed with human researchers as inventors. The US Copyright Office has similarly ruled that works created entirely by AI, without human creative input, cannot be copyrighted. In a 2023 policy statement, the USPTO clarified that "While AI systems can be tools used by human inventors, the inventorship analysis must focus on the human contributions." This highlights the crucial distinction between AI as a tool versus AI as an autonomous creator.
Lessons Learned§
The consistent outcome across multiple jurisdictions—Japan, the US, the UK, and the EU—demonstrates a global consensus that patent law is fundamentally rooted in human agency. The DABUS campaign, while failing to achieve its goal of AI inventorship, has nonetheless stimulated important legal and philosophical debates about the nature of innovation and the role of AI. As one commentator insightfully noted, "If an AI can invent something then it should be considered obvious"—a nod to the legal principle that patents require non-obviousness, which may be undermined if AI can routinely produce novel inventions. The counterargument, that AI might enable unprecedented levels of innovation, remains a challenge for future legislatures to address. For now, however, the legal framework remains clear: only humans can be inventors, and those who attempt to bypass this requirement are likely to face rejection.