Acceleration of Scientific Research by AI and Intellectual Property Systems in the Era of Superintelligence (Need for Superintelligence Legal Science)

A paper about The AI Scientist was published in August this year (Chris Lu, Cong Lu, Robert Tjarko Lange, Jakob Foerster, Jeff Clune, David Ha: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery, arXiv:2408.06292 [cs.AI](2024)).

The AI Scientist can automatically perform scientific research.　If AI is capable of doing research, AI can do AI research, and an era of superintelligence is expected to come that is far beyond human intelligence.

In the past, people said that AI could not perform creative tasks. However, due to the advancement of AI technology, AI has been able to make inventions.

The JPO published an analysis about AI inventions in April this year.  The JPO decided to monitor the further progress of AI-related technologies.

Only four months later, in August of this year, The AI scientist was published, and rapid technological development became evident.  It is now time to consider the patent system for AI inventions.

In the era of superintelligence, there are many points to be improved in the Japanese intellectual property system. For example, there is a pressing need for a fundamental solution to the issues of generative AI and copyright. Japan introduced Article 30-4 of the Copyright Act. This is a creative system that was first introduced in Japan (social innovation). However, Article 30-4 of the Copyright Act could not fundamentally resolve the issues of generative AI and copyright and it is still therefore necessary to introduce a system to fundamentally resolve the problems of generative AI and copyright.

There is a great need for companies to introduce generative AI. When generative AI further develops in the future, productivity improvements will be enormous with the introduction of generative AI. Whether or not the issues of generative AI and copyright are fundamentally resolved will have a major impact on the productivity of many companies. Thus, resolving the issues of generative AI and copyright will be extremely important.

Moreover, it is crucial to introduce a new intellectual property system for data. Currently, the Unfair Competition Prevention Law in Japan has a system of “Gentei Teikyo Data (limited provision data)” that protects shared data with limited access. This is also a creative system first introduced in Japan (social innovation). However, in the era of superintelligence, large amounts of publicly available AI learning data will also be needed. In addition to improving AI performance, it will also be necessary to collect data about human social norms in order to prevent illegal acts of AI.

In order to resolve these problems, it will be necessary to consider introducing the system of “Data Income (DI)” as an intellectual property system for data. The intellectual property system of the Data Income (DI) will not focus on exclusive rights, but will instead focus on collecting AI learning data from society and making effective use of such data. However, the system where everybody can possess intellectual property rights to data will also provide a perspective to prevent the Japanese Government from monopolizing and abusing data.

Thus, in the era of superintelligence, it will be essential to have new perspectives on the intellectual property system. International harmonization is necessary, but in an area where harmonization is not particularly needed, it is possible to introduce a creative system for the first time in the world (social innovation).

The era of superintelligence is rapidly emerging, and it is expected to arrive within several years. The 27th meeting of SIG-AGI in the Japanese Society for Artificial Intelligence discussed this issue.

In preparation for the rapid advent of the era of superintelligence, now is the time to begin discussions on new intellectual property systems.