Open-Source AI as Soft Power

Updated: 2025.10.13 8D ago 6 sources
China can gain leverage by exporting open-source AI stacks and the standards that come with them, much like the U.S. did with TCP/IP. If widely adopted, these technical defaults become governance defaults, granting agenda-setting power over safety norms, interfaces, and compliance. — This reframes AI governance as a standards competition where code distribution determines geopolitical influence.

Sources

China Is Shipping More Open AI Models Than US Rivals as Tech Competition Shifts
msmash 2025.10.13 86% relevant
The article reports China 'now produce[s] most of the world's freely available AI models,' with DeepSeek leading on Hugging Face and Alibaba outscoring OpenAI/Meta on LMArena blind tests—evidence that exporting open stacks is bolstering China’s influence and developer mindshare.
A Third Path For AI Beyond The US-China Binary
Dang Nguyen 2025.09.16 62% relevant
Vietnam’s FPT proposes opening its core AI stack (models, cloud, data) and a national sandbox to enable local builders, echoing the notion that open stacks set norms and governance defaults—here used by a mid‑tier country to avoid dependence on either U.S. proprietary platforms or China’s centralized systems.
UAE Lab Releases Open-Source Model to Rival China's DeepSeek
EditorDavid 2025.09.13 78% relevant
The UAE’s Institute of Foundation Models is open‑sourcing not only its K2 Think model but also the full training stack (code, datasets, checkpoints), mirroring the thesis that states can project influence by exporting open AI standards and stacks to shape norms and ecosystems.
'China Inside': How Chinese EV Tech Is Reshaping Global Auto Design
msmash 2025.09.11 48% relevant
As with AI stacks, China is exporting EV platforms and software (SAIC to Audi; GAC/Toyota; Xpeng/VW; CATL chassis), positioning its technology as de facto standards that others build on—standard-setting as geopolitical leverage outside AI.
August 2025 Digest
Thomas des Garets Geddes 2025.09.07 70% relevant
Liu Jia’s warning that advanced AI may develop 'sovereign‑consciousness' aligns with the notion that AI stacks embed governance norms; both frames treat models/standards as vehicles for geopolitical influence and argue choices about AI ecosystems determine who sets values.
Going Global: China’s AI Strategy for Technology, Open Source, Standards and Talent — By Liu Shaoshan
Thomas des Garets Geddes 2025.07.10 100% relevant
Liu Shaoshan’s roadmap emphasizes 'open‑source ecosystem development' and 'international standard‑setting' as levers for Chinese AI to 'go global.'
← Back to All Ideas