Japan amends its copyright legislation to meet future demands in AI and Big Data
Mon, Sep 03, 2018
In May this year, Japan’s National Diet approved a legislation for updating its “Copyright Act”, a reform that focused on allowing much-needed flexibility and legal certainty for innovators. The main objective of this amendment was to promote innovative digital and Artificial Intelligence (AI) services that are emerging or will emerge in the future, primarily by removing ambiguity for using copyrighted works for understanding and analysis.
In other words, Japan has insured that copyright cannot be an obstacle for AI. As the European Parliament prepares to review and vote once more on its own copyright rules on 12 September, it is worth taking a look at how Japanese lawmakers rightfully connected machine learning technologies with the development of AI, and successfully managed to balance their copyright rules to support their technological ambitions.
AI is not a buzzword in Japan: the country has long looked towards robotics and AI as solutions to help address its societal and economic issues. Prime Minister Shinzo Abe specifically put the Internet of Things (IoT), Big Data, robotics, and AI at the heart of its growth strategy, with the objective of making Japan a global leader in these fields. He is also the one who gave the political impetus and launched Japan’s Artificial Intelligence Technology Strategy Council in April 2016 to develop a roadmap for the development and commercialisation of AI in the country.
Japan also decided to step up its efforts in those areas because it is facing serious competition from its neighbour, China. The Chinese government is not hiding its ambitions of becoming the world leader on AI. It is aiming to reach parity with the US in this field by 2020, and to become the world’s primary AI innovation centre by 2030. They are well on track to achieving this by massively investing in the development of these technologies, and by offering a welcoming legal environment to data analytics firms.
Japan adopted a similar approach, and ensured its legislative environment was fully supporting its AI ambitions. That meant allowing researchers and private companies to carry out machine learning activities through an update to its copyright rules.
Japan’s copyright laws have permitted machine learning techniques since 2009. It was the first country in the world to update its copyright laws to enable text and data mining (TDM)by introducing a new Article 47(7), which authorised TDM by all users and for all purposes (commercial and non-commercial).
However, many stakeholders still found this exception was including a lot of legal uncertainties, especially around reproduction, use of databases and storage of works used for AI purposes. Similarly, one of the main criticisms to the text was that it was going to be quickly outdated. And with the rapid growth of IoT, AI, Big Data and Robotics, it had to be amended in order not to become obsolete.
The 2018 Amendment to the Copyright Act aimed to fix these issues, as well as support Prime Minister Abe’s ambitions, and his objective to promote the development of Japanese AI and Big Data industries, two technologies that heavily if not completely rely on Text and Data Mining. The 2018 Amendment introduced the following three provisions, removing perceived copyright barriers to AI:
- New article 30-4 lets all users analyse and understand copyrighted works for machine learning. This means accessing data or information in a form where the copyrighted expression of the works is not perceived by the user and would therefore not cause any harm to the rights holders. This includes raw data that is fed into a computer programme to carry out deep learning activities, forming the basis of Artificial Intelligence;
- New article 47-4 permits electronic incidental copies of works, recognizing that this process is necessary to carry out machine learning activities but does not harm copyright owners;
- New article 47-5 allows the use of copyrighted works for data verification when conducting research, recognizing that such use is important to researchers and is not detrimental to rights holders. This article enables searchable databases, which are necessary to carry out data verification of the results and insights obtained through TDM.
The Copyright Amendment Bill will come into force on 1st January 2019 and is expected to accelerate technological developments in fields of critical economic importance for the country.
Japan is just one example of a country that has taken the necessary steps to create the best legal environment to encourage the development of future technologies, which will have a critical role in guaranteeing global economic growth in the next years and decades.
As Japan is moving forward, we urge the European Parliament not to go three steps backwards when the rest of the developed world is walking strides ahead in the opposite direction. We call on MEPs to adopt future-proof copyright rules that encourage innovation on 12th September.
 For the official texts (in Japanese): http://www.mext.go.jp/b_menu/houan/kakutei/detail/1405213.htm
 For more information on the 2009 text : https://www.futuretdm.eu/wp-content/uploads/FutureTDM_D3.3-Baseline-Report-of-Policies-and-Barriers-of-TDM-in-Europe-1.pdf
 Accenture report, Why AI is the Future of Growth: https://www.accenture.com/t20170927T080049Z__w__/us-en/_acnmedia/PDF-33/Accenture-Why-AI-is-the-Future-of-Growth.PDFla=en
Other blog posts
The European Alliance for Research Excellence: from text and data mining to open data
Tue, Oct 13, 2020
Read EARE's new focus on open data now that the text and data mining and Copyright Directive discussions are over.
LIBER joins the European Alliance for Research Excellence
Tue, Apr 20, 2021
The European Alliance for Research Excellence (EARE) is thrilled to welcome LIBER as the newest member of its coalition.
Data Act: Creating a coherent framework to increase stakeholders’ trust in data sharing
Tue, Sep 07, 2021
Read our contribution to the European Commission's consultation on the Data Act.