Skip to content
note·Indexed by this site · Originally published· Governance →· Its place in the scan record →

IISE researcher compares the two camps of the Hiroshima AI Process and WAICO and analyzes risks for Japanese companies

IISE研究员讨论国际AI安全治理与依赖风险

AI summary

Yuji Shimoyama, a senior researcher at the Institute for International Socio-Economic Studies (IISE), published an article comparing the two major international AI governance frameworks—the Hiroshima AI Process (HAIP), in which Japan participates, and the China-led WAICO—and analyzing dependency risks and implications for Japanese companies. HAIP was established in May 2024 and, as of October 2026, has 69 countries and the EU participating, with companies such as Anthropic, OpenAI and NEC participating in its reporting framework; WAICO was proposed by China and established in July 2026 and, as of August 2026, has 37 countries participating, 12 of which overlap with HAIP. Using IMF nominal GDP, the author estimates that in 10 years the HAIP camp will still hold an economic advantage. On risks, the article says the United States once asked OpenAI and Anthropic to suspend providing models to the UK AISI for pre-deployment review, and that this reportedly actually stopped for 17 days; the author also worries that China, through capacity-building assistance, will bring about de facto vendor lock-in for Chinese-made AI in important markets for Japanese companies such as Vietnam, Indonesia and Malaysia.

Read originalnote.com