The Yomiuri Shimbun

Japan is launching a program to support domestic companies and other entities in large-scale data collection for AI training to accelerate the development of artificial intelligence used to operate robots and other systems, according to sources.

The Economy, Trade and Industry Ministry has selected nine data collection projects--covering fields such as health care, household services and infrastructure--to receive support, providing a total of up to 14 billion yen in subsidies.

Physical AI, which acts as the "brain" for robots and other devices, requires training data such as on the operation of equipment used in various settings and human movements in order to improve its performance. By supporting the collection of such data, the government aims to create an environment that promotes the development of domestically produced AI.

Examples of eligible projects include: Kobe University's collection of data on surgical robots; operational data from robots deployed in homes within Toyota Motor Corp.'s "Woven City" demonstration site by the AI Robot Association; operational data from staff at the housekeeping service Bears Co.; and robot operation data collected by major manufacturers Kawasaki Heavy Industries Ltd., Yaskawa Electric Corp. and Fanuc Corp.

This will help to improve the performance of AI being developed domestically by Noetra Corp., a company established by a consortium of firms including SoftBank Corp.

While the government also envisions providing data to other AI development companies, it will impose certain conditions on providing data to overseas firms.

Publicly available data for AI learning on the internet is said to be mostly used up, making it increasingly important for companies and other entities to collect data independently.

The economy ministry aims to enhance AI performance using data collected through this support program, thereby promoting the widespread adoption of AI and collecting even more data.

----

This article is from The Yomiuri Shimbun. Neither Dow Jones Newswires, MarketWatch, Barron's nor The Wall Street Journal were involved in the creation of this content.

YDN-M0000214177-1