Global AI Guidelines: Formulate Regulations by Putting Top Priority on Safety

Editorial

15:36 JST, December 8, 2023

It is significant that the Group of Seven advanced nations have compiled international rules toward regulating artificial intelligence, which has been pointed out to have various risks, including misuse in crime.

The Japanese government needs to take effective measures.

The G7 leaders held an online summit to compile comprehensive international guidelines on AI.

The guidelines include the establishment of a mechanism for experts to inspect the safety of content created with advanced AI systems and the need to introduce technology to distinguish content created by humans from AI-generated content.

Major U.S. tech companies are developing one new AI model after another. As more diverse AI models are introduced into the market, there is a greater risk that they will be misused. Among the risks pointed out are that AI models could teach people how to make weapons or spread discriminatory ideas.

It is essential for experts in each country to take responsibility to investigate whether there is a risk of AI models disrupting society, before they become widely available.

In Europe, laws are being established to create regulations such as requiring developers to clearly state if content was AI-generated. Japan, on the other hand, is prioritizing the development and spread of AI models and intends not to go beyond self-regulation, based on the assumption that developers and users will abide by rules.

However, Japan has already seen the proliferation of sophisticated fake videos in which images of politicians, such as Prime Minister Fumio Kishida, appear to be speaking, stirring up controversy in society.

Is the government going to leave the situation as it is? Shouldn’t it establish technology that can tell whether AI-generated content is real or false information, and legally regulate the dissemination of false information?

Behind the ease of creating fake videos with AI is the fact that the government amended the Copyright Law in 2018 to allow AI models to learn from copyrighted materials without seeking permission.

It is of the utmost necessity to amend the Copyright Law again to establish a new system so that copyrighted works can be excluded from materials used to teach AI if copyright holders wish to do so.

Also on the agenda at the summit were the situations in Ukraine and the Middle East.

In response to Ukrainian President Volodymyr Zelenskyy’s appeal for continued support, Kishida announced that Japan would provide $1 billion in additional aid for the restoration and reconstruction of infrastructure and other elements in Ukraine. It is hoped that Japan, which cannot provide military assistance like Western nations, will continue to cooperate through the civilian sector.

Japan has served as G7 chair this year. Although its role has ended with this summit, it does not change the fact that Japan still has a responsibility to take the lead in rule-making in the international community as a representative of Asia. Tokyo should make every possible diplomatic effort to restore the world order.

(From The Yomiuri Shimbun, Dec. 8, 2023)