Japan has made its ruling on the situation between Content creators and Businesses. Japanese companies that use AI have the freedom to use content for training purposes without the burden of copyright laws.
This news about the copyright laws in Japan reported over at Technomancers is seen as Businesses: 1 / Content Creators: 0
The Japanese government confirmed that it will not enforce copyrights on data used in AI training. This will open the door for AI systems to make use of any data, regardless of its origin or purpose. As both companies and countries are realising the importance of the AI industry, this decision adds an additional dimension to debates on AI regulation. Japan's move is inline with its motivation to become a leader in AI technology.
Japan clearly plans to avoid any hindrance to AI research and development. And sees this as a way to better leverage this technology and compete directly with the rest of the world.
This follows a similar decision from Israel in January whereby AI based solutions are exempt from legal matters of copyrighting for training. It describes how the “value of AI systems depends first and foremost on the quantity of the materials that the machine is fed at the ML stage, together with the diversity and quality of such materials.”. That “given the immense quantity of works that ML entails, the ML process will almost always require the use of works that are copyrighted by multitudes of third parties.”. Also “lifting the copyright uncertainties that surround this issue can spur innovation and maximize the competitiveness of Israeli-based enterprises in both ML and content creation.”
Like Japan, they have taken a very simplistic, and opportunistic view of the value of copyrighted content. The discussion about copyright laws has already filled up books and articles, but the idea is that they protect works of value. This is understandably thought of as being on a different level to AI, which is predicted to have a revolutionary effect on society. However, taking such a position then means that the resulting AI systems should also be void of both copyright coverage. The resulting AI models should therefore be available for open source solutions to use them for training.
The disclosure of data used for training, and getting to the core of this topic, explaining and divulging which copyrighted materials were used should be regulated. But such regulations should not just strip AI tools and systems of all and any obligation to do so and provide transparency for researchers to better understand them, track sources of bias etc. To also recognise that AI tools are a reflection of the data they use as their footing. Copyrighted or not.