I'm going to stop because there is no end in sight, but it is finally becoming a matter of course to implementing AI－generated products.
Now, then, what should be done about the rights there...
It is time to get serious about it.
In the previous IP promotion plan
https://www.kantei.go.jp/jp/singi/titeki2/tyousakai/kensho_hyoka_kikaku/2017/johozai/dai6/siryou4.pdf Secretariat for Intellectual Property Strategy Promotion, Cabinet Office, Government of Japan, "Remaining Issues Concerning AI ( (for discussion)," February 28, 2009, p.2
https://www.kantei.go.jp/jp/singi/titeki2/tyousakai/kensho_hyoka_kikaku/2017/johozai/dai6/siryou4.pdf
Secretariat of Intellectual Property Strategy Headquarters, Cabinet Office, Government of Japan, "Remaining Issues Concerning AI (for Discussion)," February 28, 2017, p.2.
The basic idea is that AI is a "good" thing.
So ・・・・
------------------------------------------------------------------------
The biggest difference between AI-generated products and computer-generated products is that according to the Intellectual Property Promotion Plan 2017 (https://www.kantei.go.jp/jp/singi/titeki2/kettei/chizaikeikaku20170516.pdf), AI-generated products are "human creative The biggest difference between AI-generated works and computer-generated works is that AI-generated works are considered "autonomously generated by AI without any creative contribution by humans," according to the IP Promotion Plan 2017 (), in which case they do not have copyrightable characteristics. On the other hand, "With respect to creations utilizing AI, if there is a human creative contribution in the process of creating an AI-generated product to obtain an AI-generated product, which is a concrete output, it is considered that AI was used as a 'tool' and the AI-generated product is copyrightable (Intellectual Property Plan 2017, p. 13)" and that human If there is a creative contribution, the work is considered to be copyrightable; whether or not it is an AI-generated work depends on whether or not there was a creative contribution by a human being.
---------------------------------------------------------------------
If we go by this, there is no right for AI unmodified products.
However, if some creative modification has been made, then the person who made the creative modification is considered to be the creator of the work.
However, can this boundary be easily divided?
Let's divide it simply into the following cases.
Classification of cases
Wouldn't it be something like this?
In short, it is difficult to judge, and in the case of AI, it is also difficult to determine who was involved in the creative process.
My own theory is the creation theory (?). I once thought that creativity should be the key point, and that it would be better to judge the use of AI as a promotion scheme without considering the use of AI itself as a problem, but rather by determining where the creativity came from.
Scheme.
In essence, it would be the same as control point judgments
Judgment scheme
There are criticisms of control point judgments
Copyright in AI-generated Works: Lessons from Recent Developments in Patent Law by Rita Matulionyte, Jyh-An Lee :: SSRN
The point is that the owner of AI should be attributed the rights, but this may be a problem when the user and the owner are different as in this case (this is also acknowledged in the paper).
If it is not the owner of the AI, then
(1) AI itself, (2) programmers, (3) users, (4) public domain, (5) key actors, (6) those who gave creativity, and (7) co-authorship by related parties. (7) The parties concerned become co-authors.
(There are papers that argue for each of these, but I will omit them. (There are papers arguing for each of these, but I won't go into them here. (This is a personal memo blog, so it is simple.)
However, in any case, there would be problems in terms of complexity of AI generation and incentive theory.
Incidentally, since (5) and (6) above can be summarized as almost flexible incentives (with some minor differences), I have combined them into one as (5).
Evaluate incentives based on 7 categories of various theories
|
number |
method |
Incentive Evaluation |
supplement |
|
① |
When attributing to AI |
No incentive for AI itself (as of now) = Cannot be evaluated from an incentive theory perspective. |
Need to discuss legal reform with AI as a whole |
|
② |
Attributed to the programmer and company that created the AI |
Incentives are created for AI creators, but it is assumed that users will avoid using the AI, and as a result, it is difficult to evaluate. |
Response to GitHub Copilot |
|
③ |
Belonging to the user |
Suitable, but may lead to mass creation generation by copywrite troll-like AI and abusive lawsuits if even remotely similar to it. Careful institutional design is necessary. |
Olympic Emblem Riots |
|
④ |
Public domain (no attribution) |
The use of AI means that copyright cannot be obtained and cannot be evaluated from an incentive-theoretic standpoint. Also note the subsequent transaction costs. |
See also the argument by the argument that copyright is unnecessary. |
|
⑤ |
Flexible attribution A To the pivotal actors |
Optimal, but difficult as decisions are made on a case-by-case basis and disputes are more likely to arise. |
In a sense, we leave it to case law. |
|
⑥ |
Flexible attribution B To the Creator |
Ditto and most faithful legal approach. Attribution method Currently in place. But not sure if it necessarily fits in the AI era. |
Faithful to copyright law |
|
⑦ |
The parties involved are co-right holders |
Increased complexity of rights relationships and the increase in the number of rights holders who are not necessarily friends cannot be evaluated from an incentive theory perspective. Transaction costs also increase dramatically. |
The difficulty of joint rights |
Incentive Comparison
I may not have said it well enough, but I think I can summarize it simply as follows
(This is a very brief summary)
A Study on the Attribution of Rights to AI Products - Focusing on the Motivation of Actions - (jst.go.jp)
Also, already out of stock
The World of Intellectual Property Jurisprudence - In Commemoration of the 10th Anniversary of the Graduate School of Intellectual Property Studies, Kibi International University, Essays in Commemoration of Teruo Doi and Shinichi Kuzuminato's Yomei Memorial Editions | Kibi International University Graduate School of Intellectual Property Studies Editorial Office | Books | Mail Order | Amazon
(I have taken various things from the papers contained in this book.)
So, the other is the secondary work composition.
This is a method of composing an AI-generated work as a secondary work rather than the original.
However, there are many problems to be solved, such as how to connect primary authors and secondary authors.
We are thinking of putting a group in charge of this.
The picture would look something like this...
Conceived
Well, I wonder about the validity...
(By the way, I actually brought this from the PPT of the presentation at the International Conference on Public Economics...)
I have written too much about my own ideas at this point.
Of course, there is a need for further deep consideration....
The above is just a personal memorandum.
Here is a miscellaneous summary of the government-related activities.
--------------------------------------------------------------
There are various parallel developments in intellectual property law regarding this matter, and a great deal of consideration is being given to it. For example, the Interim Summary of the Subcommittee on Legislation and Fundamental Issues of the Copyright Subcommittee of the Council for Cultural Affairs (February 2017 http://www.bunka.go.jp/seisaku/bunkashingikai/chosakuken/pdf/h2902_chukanmatome.pdf) states that "(omitted) (2) Specifically, with technological innovations such as IOT, big data, and artificial intelligence, as well as the accumulation, processing, and dissemination of information becoming easier and cheaper, the creation of new innovations that create added value by accumulating, combining, and analyzing large amounts of information is expected, and discussions at the government's Intellectual Property Strategy Headquarters have also promoted this. In the discussion at the government's Intellectual Property Strategy Headquarters, the necessity of building a next-generation IP system to promote this and to link it to the development of a new culture that enriches society was stated (Interim Summary of the Subcommittee on Legislation and Basic Issues, Copyright Subcommittee, Council for Cultural Affairs, p. 4). (Interim Summary of the Subcommittee on Legislation and Basic Issues, Subcommittee on Copyright, Council for Cultural Affairs, p. 4). In addition, the report of the "Study Group on the Intellectual Property System with a View to the Fourth Industrial Revolution" published by the Ministry of Economy, Trade and Industry on the specific form of protection for AI-based creations (http://www.meti.go.jp/report/whitepaper/data/pdf/ 20170419001-1.pdf) states, "(Omitted) The scope of data utilization is limited, partly because social understanding of data utilization has not been sufficiently fostered. In particular, in the case of data related to manufacturing generated in factories, data utilization among companies is extremely cautious, as it may contain ideas and know-how directly related to competitiveness. In considering the protection of such data, data that has emerged as a result of technological progress, such as data for learning that would allow AI to learn, data prior to the construction of this learning data, and parameters contained in learned models after AI becomes smarter as it acquires data, respectively, should be protected. Data that has emerged as a result of technological development, such as data before the construction of this learning data, and parameters contained in learned models after AI has become smarter by acquiring data, can be protected as copyrighted works if their creativity is recognized, as patents if they meet the definition of invention, and as trade secrets if they meet the three requirements of (1) confidentiality, (2) usefulness, and (3) non-public knowledge. (pp.3-4, "The IP System in the Context of the Fourth Industrial Revolution").
----------------------------------------------------------------------
Secretariat of Intellectual Property Strategy Headquarters, Cabinet Secretariat, "Handling of Creative Works Produced by AI (for Discussion)," 2016 (This document is used hereafter as a source showing the main views of the Secretariat of Intellectual Property Strategy Headquarters, Cabinet Secretariat. http://www.kantei.go.jp/jp/singi/titeki2/ tyousakai/kensho_hyoka_kikaku/2016/jisedai_tizai/dai4/siryou2.pdfp16.3-1 slide), two problems are considered if AI-generated works are given protection equivalent to copyright First, "the right to create" is not the same as "the right to use". First, "the number of creations with rights will increase explosively," and second, "the subject of rights is ambiguous (based on the philosophy of the copyright system, "the person who made a creative contribution" would be the subject of rights, but this cannot be specified objectively because it is determined by the actual state of creation). The latter is the problem of authorship. The latter is the issue of authorship.
The above slide of the Secretariat of Intellectual Property Strategy Headquarters, Cabinet Secretariat states, "Candidates for the subject of the right may be (1) the program developer, (2) the provider of data for learning artificial intelligence, and (3) the person who instructed the artificial intelligence with the intention of creation, but how should they be evaluated? (Secretariat for the Promotion of Intellectual Property Strategy, Cabinet Secretariat, above, 2016 p.18.3-3 slides)
------------------------------------------------------------------------------------
The following was shown.
**-----------------------------
The ones circled in * are straight from old explanations, a bit out of date.
Now, the government system should be moving on a lot soon, so we need to watch for that as well....
