The Ministry of Science and ICT and the Korea Radio Promotion Association have decided to provide KRW 4.83 billion each to four companies—KBS, MBC, regional MBC, and KT ENA—to build AI training data based on Korean broadcast content. These broadcasters plan to build a total of 21,000 hours of training video data.

On the 8th, the Ministry of Science and ICT announced that it had selected a consortium of four companies—KBS, MBC, regional MBC, and KT ENA—as recipients of the broadcast-video AI training data construction project. These broadcasters will build AI training data using original domestic broadcast footage with copyright issues resolved. The Naver Cloud, Upstage, SK Telecom, NC AI, and LG Management Development Institute AI Research Institute consortium (the national representative AI team), selected for the “Independent AI Foundation Model Project,” may request and use this data, and some data for AI development for research and educational purposes will also be opened.

A consultative body for data sales will also be operated. Broadcasters, AI and data experts, and companies demanding data are expected to participate and discuss standards for trading AI training data. The Ministry of Science and ICT will provide KRW 4.83 billion to each broadcaster, for a total of KRW 19.32 billion.

MBC formed a consortium with iMBC, CrowdWorks, DataMaker, and the LG Management Development Institute AI Research Institute. MBC plans to build AI training data that generates high-precision, high-quality background videos using news, current affairs/culture, entertainment, and drama footage. Fifteen regional MBC stations, including MBC Chungbuk, together with Dost11, GeminiSoft, the Electronics and Telecommunications Research Institute, and Ufit, will build image and video generation data reflecting the culture and lifestyles of Korea’s regions.

Source: Media Today (https://www.mediatoday.co.kr)