Full-Scale Effort to Secure Data for AI Foundation Model Development
Incentives Offered, Including Data Processing Support and Priority Access to Services
The Ministry of Science and ICT announced on June 23 that it will begin collecting large-scale training data from public and private data-holding institutions to support the development of a domestic artificial intelligence (AI) foundation model.
Starting today and continuing until July 21, the Ministry will conduct an open call for data providers as part of the "Independent AI Foundation Model Project." The goal is to develop a world-class large language model (LLM).
This project is a key AI strategy initiative that the government is actively pursuing from this year. The plan is to provide selected development teams?chosen through future calls for proposals?with concentrated support for core infrastructure such as GPUs, data, and talent, thereby enabling the development of an independent large-scale AI model.
Eligible data providers include public and private institutions that possess high-quality data, whether paid or free. The final use of the data will be determined by agreement after the selection of the development teams, based on their needs. Agreements will be signed among the data providers, the National Information Society Agency (NIA) and Korea Data Agency (K-Data) under the Ministry of Science and ICT, and the development teams.
The government will offer various incentives to data providers. These include support for data processing and preprocessing costs, and reimbursement for the use of paid copyrighted materials that meet certain conditions. Additional incentives under consideration include priority access to AI services, extra points in government projects such as data vouchers, and support for distribution and transactions through AI Hub.
Kim Kyungman, Director-General for AI-Based Policy at the Ministry of Science and ICT, stated, "High-quality data is a core element that determines the performance of artificial intelligence," and added, "A data collaboration system jointly built by the public and private sectors will lay the foundation for the development of a world-class Korean LLM."
After recruiting data providers, the government plans to select development teams in August and sign data agreements based on demand. The government intends to clearly define data provision conditions, scope of use, costs, and responsibilities in order to minimize potential disputes. Applications from data providers can be submitted via email, and must include an application form, data provision plan, and business registration certificate, among other documents.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
