Connect with us

Economy & Tech

EDITORIAL | AI Safety: Guard Against Risks Outweighing the Benefits

The first multilateral discussion on AI safety called for international cooperation while identifying concerns about AI's risks if used to gain hegemony.



British Prime Minister Rishi Sunak and other dignitaries attending the AI Safety Summit. ©REUTERS/Toby Melville/Pool via Kyodo)

The two-day AI Safety Summit, hosted by the United Kingdom and held near London, wound up on November 1. It was the world's first multinational effort focused on such things as risk management for generative AI. The technology, which includes ChatGPT, is undergoing tremendous development. 

ChatGPT performs intelligent processing through natural dialogue with human beings. 

Discussions among government officials and representatives of AI companies from 28 countries, including Japan, the UK, the United States, and the European Union, resulted in the "Bletchley Declaration" on AI Safety. It is named after Bletchley Park where the meeting was held. 

The declaration focused on "sharing an international understanding of the risks and possibilities associated with AI." Among other things, it includes state-directed validation and evaluation of next-generation AI models prior to their public release by developers. 

That should be considered a positive outcome. There were fears that a lack of transparency in development could lead to catastrophic damage from biological and chemical weapons, cyberterrorism, and disinformation

Yi Zeng from the Chinese Academy of Sciences speaks ahead of the welcome on Day 1 of the AI Safety Summit at Bletchley Park on November 1, 2023. (©Leon Neal/Pool via REUTERS)

Different Perspectives on Regulating AI

Nonetheless, many problems remain to be resolved. Different countries look at generative AI in different ways. 

The differences are particularly striking with respect to regulation. For example, the EU is looking to prevent AI abuses through strict laws, including huge fines. Meanwhile, the UK is emphasizing the use of AI for economic growth and is reluctant to impose tough regulations. Japan is also inclined to follow a moderate approach relying on guidelines rather than laws and regulations that may hinder innovation. 

On the eve of the AI Safety Summit on October 30, US President Joe Biden signed an executive order on AI regulation. It seeks to promote development while largely leaving it up to the AI development companies to self-regulate. 

Although China participated in the summit, it is currently tightening its arbitrary controls on the Chinese people. It seems unlikely that it will agree to work in concert with democratic countries. 

Luminaries rush to their seats for a plenary session at the AI Safety Summit at Bletchley Park, Britain, November 2, 2023.  (©Alastair Grant/Pool via REUTERS)

Varied Hopes and Expectations

This first multilateral discussion on AI safety called for international cooperation. However, it was very much a case of the various participants having their own hopes and agendas for AI. Each was eager to seize the initiative to create rules favorable to the development of AI in their own country. 

Japan was a latecomer in terms of AI technological development. However, as the host of the G7 Summit in May, it has led the "Hiroshima AI Process." The process is a framework for establishing international rules for the use of AI. The public and private sectors need to concentrate their efforts without delay. 


Artificial intelligence, or AI, which could eventually surpass human intelligence, is an unprecedented invention born of advanced technology. If used as a means to gain hegemony, the risks will quickly outweigh the benefits.


(Read the editorial in Japanese.)

Author: Editorial Board, The Sankei Shimbun

Our Partners