Dokpamo (Sovereign AI Foundation Model)
Dokpamo
Definition
Korean abbreviation of "독자 AI 파운데이션 모델" (independent AI foundation model): a large-scale AI base model that a country builds and controls on its own, free of dependence on foreign vendors or licenses. It is the centerpiece of Korea's sovereign-AI strategy.
Examples
- • The government spent ₩530 billion to select 'national champion' teams to build a dokpamo.
- • To cut reliance on foreign models, calls grew for a sovereign AI like the dokpamo.
Origin
Spread in 2025 as the short form of Korea's government-funded "independent AI foundation model" program (Ministry of Science and ICT / NIPA).
A "foundation model" is a large AI trained on vast data that serves as the base for many tasks. Dokpamo prepends "독자적" ("one's own / independent"), stressing that the model is built, owned, and controlled by the nation itself rather than by a foreign company.
The underlying idea is sovereign AI — securing AI capability across data, infrastructure, models, and applications at the national level, without depending on a specific overseas vendor or license structure.
In Korea, the Ministry of Science and ICT and NIPA back competing "national champion AI" teams with a ₩530 billion program (selecting five teams in August 2025 and narrowing them through staged evaluations), releasing results as open weights. The goal is a Korean AI model that ranks among the world's best.