“Intensifying geopolitical competition leaves AI bridge powers in a difficult situation where they’ll soon likely face insurmountable barriers to independent frontier AI development. To stay relevant and thrive economically, they need to work together and strategically choose their AI development approaches.” – Professor Yoshua Bengio, AI pioneer and co-author of the report.
KAIST (President Kwang Hyung Lee) announced on January18th that a research team led by Professor Kyung Ryul Park from the Global Center for Development and Strategy (G-CODEs), in collaboration with Canada's Mila - Quebec AI Institute, Oxford University, RWTH Aachen University, and other global institutions, has co-published a policy report titled 「A Blueprint for Multinational Advanced AI Development」.
The report underscores that “an international advanced AI research & development partnership of AI bridge powers can feasibly produce frontier AI models, and is essential for safeguarding the sovereignty, democratic values, economic competitiveness and growth, technical innovation, and national security of these bridge power states.”
The report highlights that approximately 90% of global AI computing capacity is concentrated in the U.S. (75%) and China (15%). It analyzes that this concentration of resources could restrict "bridge power" nations from independently developing advanced AI and deepen technological dependence on specific countries or global Big Tech companies.
According to the report, "AI Bridge Power Nations" refer to countries that, while not superpowers like the U.S. or China, possess world-class research influence and digital infrastructure, but face realistic constraints in building hyper-[K글1] scale AI systems alone. South Korea, Canada, the UK, Germany, and Singapore are cited as representative examples.
South Korea, in particular, possesses strong government commitment and excellent ICT infrastructure. However, it faces limitations compared to the U.S. and China in securing hyper-scale AI infrastructure. This context aligns with the Korean government’s recently announced "AI Action Plan," which emphasizes expanding international AI cooperation.
The cooperation model proposed in the report is a multinational joint research framework similar to CERN (European Council for Nuclear Research), focusing on:
- Sharing computing infrastructure
- Cooperation on high-quality data
- Cross-border exchanges of talent and research
Through this framework, the report suggests jointly developing frontier AI models while simultaneously building an inclusive AI ecosystem that reflects ethical AI use as well as linguistic and cultural diversity. Furthermore, the report proposes strengthening the long-term technological self-reliance and innovation capacity of participating countries.
Professor Holger Hoos of RWTH Aachen University evaluated this initiative as “An ambitious but extremely realistic plan for a new multinational partnership in the field of artificial intelligence.”
Professor Kyung Ryul Park of KAIST added, “This report shows that AI bridge powers, including South Korea, can present an alternative path through scientific solidarity. It is a new opportunity for us to strengthen responsible AI leadership by leading agendas that jointly respond to global challenges.”
This report involved world-renowned scholars from Oxford University, Mila, The Future Society, and the Paris Peace Forum, with KAIST’s G-CODEs playing a central role in South Korea.