The government is moving to secure the necessary training data for the development of a 'proprietary artificial intelligence (AI) model.'
The Ministry of Science and ICT announced on the 23rd that it is publicly recruiting data supply organizations as part of its ongoing 'proprietary AI foundation model project.' If public or private organizations wish to provide data for training the proprietary AI foundation model, they can apply by July 21.
The proprietary AI foundation model project aims to build a proprietary model that has over 95% performance compared to the latest global AI models released within the past six months. The elite development team selected through the upcoming recruitment will be able to freely propose graphic processing units (GPUs), data, and talents needed for development to the government.
The data supply organizations being recruited this time will provide free or paid data that can be used for AI learning to the elite AI development team. The final utilization of the data held by each organization will be confirmed through an agreement based on actual demand after the selection of the development team. This agreement will be concluded among the Ministry of Science and ICT, the data supply organizations, and the elite development team.
Institutions that have confirmed data provision will receive financial support, including ▲ support for data processing and preprocessing expenses ▲ compensation for fees that meet the requirements when providing paid works. Opportunities to adopt ▲ AI model-based services developed through the project will be prioritized, or ▲ additional points will be considered in data voucher support projects. The Ministry of Science and ICT plans to actively support data distribution and transactions through platforms such as the AI Hub, depending on the wishes of the supply organizations.
Kim Kyung-man, the AI-based policy director at the Ministry of Science and ICT, said, "High-quality data is a key asset that determines AI performance, and a data collaboration system created by both public and private sectors will lay the foundation for the development of world-class Korean-style large language models (LLMs)." He added, "I hope that this cooperation will further expand the data alliance between the public and private sectors for the success of this project."