Intellectual Property Headquarters / New Information Goods Committee

Second meeting of the Intellectual Property Headquarters New Information Goods Committee. The proper guardianship and use of AIs. Debating institutional theories concerning training data, AI programs, and trained models.

AI is a boom and, while we are unable to anticipate a general or all-purpose AI, countless individual, single-purpose AIs are being created. The combination of these is important.

A time draws near when collective AIs, rather than the wisdom of crowds, demonstrate their power. ...this is a discussion held with this radically developing situation in mind.

Industry uses specialty AIs that bear human knowledge as a mechanism through deep learning, rather than general-use AIs.  This year’s mission is to anticipate the possibilities therein and consider institutions for using them. The topic will revolve around such things as exclusive use rights (copyrights), trade secrets, and contracts.

For example, copyright infringement of a work that was created by an AI is said to be a grey area. There are cases which, in the U.S., fall under fair use, but are illegal in Japan. Even if we were to organize a domestic system, the point of view considering international competition and usage will grow in importance.

Committee Member Miyajima indicated that, as Japan is very cautious about gray areas, it is important to examine the issue with an eye towards increasing creative work and business opportunities and to get out that message. Committee Member Yanagawa commented to note that the balance between guardianship and utility may be disrupted. I think that’s true.

Committee Member Shimizu observed that open source is mainstream in the software world and that the time of copyrights is ending. Also, that PPAP is making money on derivative works on YouTube. It’s wonderful that the name “Pikotaro” will live on in the official record of government proceedings.  

Committee Member Seo brought up the issue of a non-human “AI individual”. He suggested that we anticipate big-picture suppositions about various AIs ranging from infancy, when the AI is still learning, to adulthood. ...again, a development that indicates the situation.

Committee Member Shimizu proposed that Japan is lacking in computational resources. Therefore, we should organize an environment for everyone to use, what we might call an “AI Portal”. In response to this, there was also the suggestion that it would be good to have a “National AI Archive” much like our National Diet Library. Suddenly, we’ve jumped from institutional theory to policies for promotion.

Committee Member Kitsuregawa: This is a contest of controlling data more than it is an algorithm-like methodology.  The written works held by libraries will become less significant compared to data. We need to raise the priority on data. Right now, we need to put together a strategy for intensely concentrating resources such as data, budget, and personnel rather than the legislative system on the issue.

The work to solve this equation, which involves so many different variables including institutions, promotional policy, the advance of technology, and the international situation, will go on.