G7 sets out guiding principles for AI developers to address risks

The Group of Seven industrialized nations on Monday announced international guiding principles and a code of conduct for companies developing advanced artificial intelligence systems in a bid to address privacy concerns and the risk of misuse.

The documents aim to "promote safe, secure, and trustworthy AI worldwide," the G7 said, noting that developers, while harnessing the opportunities of innovation, should make sure that the technology is not deployed in a way that undermines democratic values, facilitates terrorism or poses "substantial risks" to human rights.

The G7, involving Japan, the United States and the European Union, has been discussing the need for international standards to harness the technology amid the rapid spread of generative AI tools, such as ChatGPT.

In a leaders' statement issued along with the principles and code of conduct documents, the G7 vowed to accelerate the process toward developing a comprehensive policy framework by the end of this year, which will include overall guiding principles for all AI actors.

As this year's chair of G7, Japan will "continue to work for international rulemaking of generative AI," Prime Minister Fumio Kishida posted on X.

In the 11-point principles and a code of conduct that laid out the details, AI developers are urged to take "appropriate measures," including independent external testing, to identify and mitigate risks.

Among the risks that warrant attention, the G7 pointed to chemical, biological and nuclear risks, cautioning that advanced AI systems can "lower barriers to entry, including for nonstate actors, for weapons development, design acquisition, or use," as well as facilitation of disinformation and threats to privacy.

Organizations are also encouraged to consider, for example, using bounty systems or prizes to encourage third parties to report vulnerability issues, and to work to develop electronic watermark tools to enable users to identify AI-generated content.

Generative AI technology can utilize vast amounts of data from the internet and other sources to generate text, images, or other media in a human-like way.

The G7 members have been working to establish rules on AI-related topics to avoid the misuse of the technology after their leaders agreed in May to launch the Hiroshima AI Process, named after the western Japanese city where the summit was held.

The EU has been moving ahead of other countries in regulating the use of AI, with the European Parliament adopting in June a draft of what it calls the world's first comprehensive AI law.

Japan plans to come up with its own domestic guidelines, based on the G7 guiding principles and code of conduct.

The G7 involves Britain, Canada, France, Germany, Italy, Japan and the United States plus the EU.