<\/p>\n
South Korea is exploring new approaches to compensate news organisations for AI training data amidst rising legal conflicts, emphasising a value-sharing model that balances industry growth and media rights.<\/p>\n<\/div>\n
South Korea is moving towards a major reckoning over who pays for the data that feeds artificial intelligence. In an interview with a group of news organisations on 8 April, Hah Jung-woo, the presidential senior secretary for AI future planning, said the government wants a settlement that leaves both AI developers and news organisations better off, rather than locking them into a zero-sum fight over copyright.<\/p>\n
The issue has already become a live legal battle abroad and at home. The New York Times sued OpenAI and Microsoft in December 2023 over the alleged unauthorised use of its reporting to train AI systems, and Korean broadcasters KBS, MBC and SBS filed their own case against OpenAI in February 2026 after earlier action against Naver over its HyperCLOVA models. The disputes have sharpened concern that AI firms are benefiting from journalism without paying for it, while publishers say their work is being used to build rival products.<\/p>\n
Hah said the government believes the answer lies in a framework that shares value, rather than relying only on upfront fees. He argued that news organisations hold data essential to building a strong domestic AI industry, but that AI capability should also help media companies produce more high-quality content more efficiently. He said officials have discussed possibilities such as reward structures tied to revenue or profit, and even fund-based models that could be matched by the state, though no formula has been decided.<\/p>\n
He also suggested that not all training data should be treated in the same way. Some online material, such as personal blogs, has unclear rights status, and he said it may be better to permit training first while reserving compensation questions for later commercial use. That approach, he said, is being considered in discussions with the science, technology and culture ministries as part of the government\u2019s broader AI action plan.<\/p>\n
Beyond copyright, Hah pointed to misinformation as another area where AI could cause harm quickly and cheaply. He said South Korea\u2019s AI Basic Act is meant to be 90% promotion and 10% safeguards, with watermarking among the minimum protections. He added that, while fake news already carries penalties, the government could consider tougher punishment when AI is used to generate or spread deceptive content.<\/p>\n
Hah also linked AI policy to demographics, saying his office was not limited to technology in a narrow sense. With population decline and ageing expected to shape Korea\u2019s future, he said AI could help care for older people and make better use of their experience, while also offering tools to offset some of the economic pressures associated with low birth rates. According to him, these long-term social questions belong in the same strategic conversation as AI, energy and science.<\/p>\n
Inspired by headline at:<\/strong> [1]<\/a><\/sup><\/p>\n