G-7 to Draw Up Int'l Guidelines on Generative AI by Year-End

Tokyo, Sept. 8 (Jiji Press)--The Group of Seven major powers have agreed to draw up international guidelines by year-end for developers of generative artificial intelligence products that automatically create texts and images. The agreement was reached at the first ministerial meeting of the Hiroshima AI Process aimed at establishing international rules on AI, which was held online Thursday. The G-7 members of Britain, Canada, France, Germany, Italy, Japan and the United States plus the European Union adopted a joint statement including the plan to create the guidelines. "We'll maximize the benefits to humankind while reducing the risks associated with AI," Japanese communications minister Takeaki Matsumoto, who chaired the meeting, told reporters after the event. The content of the statement will be reported to a teleconference of the G-7 leaders to be held this autumn. At Thursday's meeting, participants confirmed that key issues related to the development and use of AI technology are how to prevent the spread of false information and ensure transparency that makes it possible to determine whether texts, images and other content were created using AI technology. They also shared the view that generative AI tools help boost productivity, promote technological innovation and solve health, welfare and climate issues. The statement outlined 10 principles that need to be followed in developing a code of conduct for companies developing generative AI products. Developers would be requested to consider social risks before such products come into practical use and strive to identify vulnerabilities in AI systems after the practical use begins. G-7 participants also agreed to cooperate with international organizations, including the Organization for Economic Cooperation and Development, to develop technology to detect false information. The G-7 leaders agreed to create the Hiroshima AI Process at their summit in Hiroshima in May. The G-7 members will seek opinions from the business and academic sectors before releasing their view on international rules on AI at year-end. END