Prime Minister Fumio Kishida unveiled plans for the establishment of an “AI safety institute” during an expert panel meeting on artificial intelligence held on Thursday. The institute is expected to be up and running by January of next year, with the aim of ensuring the safety of AI technology. Kishida also expressed intentions to formulate guidelines for AI developers by March of next year, reflecting the growing international awareness of AI safety.
During the meeting of the AI strategy council, chaired by University of Tokyo Prof. Yutaka Matsuo, Kishida emphasized the need for an organization that can conduct research on safety evaluation methods, create standards, and carry out other relevant matters. The new institute will play a central role in providing support for tasks such as developing software for assessing AI risks, and is expected to engage in international coordination as the United States and Britain also plan to establish similar organizations.
In addition to the establishment of the AI safety institute, the government is also exploring the possibility of introducing a third-party certification system for AI developers. It is anticipated that the new institute will serve as an accreditation body for this system, further solidifying its role in ensuring the safety and reliability of AI technology.
At the meeting, draft guidelines were presented to AI developers, outlining 10 principles that they should consider in their work. These principles include “human-centeredness,” “safety,” and “transparency,” with a particular emphasis on ensuring human controllability to prevent potential infringements of rights resulting from unexpected AI behaviors.
The draft guidelines also reflect international guiding principles agreed upon by Group of Seven members based on the “Hiroshima AI Process.” In particular, they call for generative AI developers to introduce technologies for identifying AI creations and clarifying their history of information, with the consideration of new technologies such as Originator Profile, which identifies the original sources of information.
Overall, the announcement of the plans for an AI safety institute and the formulation of guidelines for AI developers reflects the government’s proactive approach to addressing the growing importance of AI safety. With the establishment of the institute and the implementation of certification systems and guiding principles, Japan aims to continue leading the way in the safe and responsible development of AI technology.
