On June 4, Japan’s Patent Office held a meeting of the Intellectual Property Committee under the Industrial Structure Council to discuss reforms to the patent system in light of inventions involving AI. As the number of intellectual property cases involving AI continues to grow, there is a pressing need to revise fundamental rules—such as the definition of an invention, the recognition of inventors, and the treatment of prior art generated by AI. This discussion marks an important first step for Japan to adapt its patent system to the “AI era.”
Can an “AI-involved invention” be considered an invention?
Under current Japanese patent law, an “invention” is defined as “a highly advanced creation of a technical idea utilizing a natural law.” However, it is not explicitly stated whether this definition applies to technologies automatically generated by AI.
The Japan Patent Office has indicated the stance that “a creation by a natural person utilizing AI qualifies as an invention.”
In other words, it is not the AI that is considered to have invented something; rather, the focus is on human-driven creativity through the use of AI.
This approach is a pragmatic one that acknowledges AI’s supportive role, so long as human involvement is ensured. It aligns with precedents and regulatory practices in both the U.S. and Europe.
Is the inventor a “human” or “AI”?
While Japanese patent law does not currently define “inventor” explicitly, in practice it is assumed to be a natural person.
The latest discussion proposed that the definition of inventor should be explicitly codified so that the active use of AI does not lead to cases where no inventor is recognized.
However, the idea of recognizing AI itself as an inventor was dismissed as potentially causing confusion, and the personality or legal subjectivity of AI was explicitly rejected.
As the scope of AI-driven creativity continues to expand, identifying the “true inventor” will become increasingly ambiguous.
Future debates will likely focus on where to draw the line—how much reliance on AI output disqualifies a human’s creativity.
Can materials or papers generated by AI be considered “prior art”?
There is also ongoing debate about whether ideas or documents generated by AI can be considered “prior art” in evaluating the novelty of subsequent inventions by others.
However, since it is difficult to prove whether something was AI-generated, the proposed direction is to exclude AI-generated content from the eligibility requirements for prior art.
Because the origin of AI-generated content tends to be ambiguous, it introduces new challenges for patent and copyright practices.
We may eventually see discussions about requiring source disclosure, or managing records of AI tool usage, to ensure the reliability and traceability of “prior art.”
The Patent Office’s latest deliberation is a significant move toward updating the legal framework in response to the expanding role of AI in creative activities.
Establishing clear criteria for “AI-assisted inventions” is essential for ensuring sound technological progress and industrial competitiveness.
At the same time, it raises fundamental questions such as “What constitutes human creativity?” and “How should we define the rights and responsibilities of inventors?”
As Japan prepares for more comprehensive institutional reforms, it will be vital to engage not only companies, but also academia and the legal community in a broad-based debate.
We, too, should closely follow the future of the patent system as it evolves alongside AI.