South Korea has passed groundbreaking laws to improve sharing of government-funded AI training data, boost data interoperability, and expedite testing processes, potentially transforming its AI ecosystem.
South Korea has taken a legal step towards making government-funded AI training data more widely reusable, in a move that could reshape how public data is shared across the country’s technology ecosystem. The National Assembly on 1 October passed amendments to the AI Basic Act, the Data Industry Act and the Special Act on Promotion of Research and Development Zones, creating what lawmakers and industry observers are calling the Three AI and Data Innovation Acts.
At the centre of the reform is a new requirement that AI training data created through state-funded projects must be linked to, or made available through, the government’s integrated training-data provision system. According to the Ministry of Science and ICT, the aim is to stop valuable datasets from remaining locked inside individual institutions once a project has ended, where they can be difficult to reuse for other AI systems.
The scale of the state’s investment helps explain the urgency. Data from the National Assembly Budget Office shows that 1.6328 trillion won was invested in the ministry’s AI training-data development programme between 2017 and July 2025. During that period, 903 kinds of data were published through AI Hub. The new framework is intended to make such material easier to circulate and reuse, rather than leaving it fragmented across separate project operators.
A second pillar of the legislation focuses on the broader data market. The amendment to the Data Industry Act gives a legal basis for an integrated data platform that will connect information on where data is held and how it can be accessed across public and private systems. Rather than moving all source data into a single warehouse, the government wants to standardise “location information” so users can search one platform and then obtain the data from the institution or company that actually controls it.
That approach reflects the current state of South Korea’s data infrastructure. The National AI Strategy Committee said in last year’s national AI action plan that the country has about 200 data platforms, but that they suffer from weak standardisation, stale data and poor interoperability. The new law gives the state more room to push public bodies and private firms to connect their systems, while also allowing technical and financial support where needed.
For IT services companies, the changes could open a significant new market. Much of the work will involve data standardisation, quality control, metadata management, cloud infrastructure, access control, security and governance. Samsung SDS is among the best placed to benefit. In April, the company launched an AI-powered parliamentary support platform after linking data from 320 institutions inside and outside the National Assembly. The system is now used by about 5,000 people affiliated with the Assembly, and Samsung SDS says it supports AI search, analysis and drafting based on data from legislative, public and academic sources.
LG CNS also has a strong position in related work. Its AI and cloud revenue in the first half of this year reached 1.6714 trillion won, accounting for about 59% of total revenue of 2.8358 trillion won. The company has already been expanding AI and data platform projects in the public and financial sectors, including work for the Ministry of Foreign Affairs, the National Police Agency, the Korea Meteorological Administration and Incheon International Airport Corp. If public-sector data integration projects grow, they are likely to align closely with LG CNS’s existing business.
KT is also building its strategy around “Data for AI”, a term it uses for the data infrastructure needed to make corporate information usable by AI systems. The company is expanding its AX platform business for public and private clients, while KT Cloud is supplying dedicated cloud services and AI infrastructure to public institutions. KT has said its 2026 AX strategy will strengthen the base for that business line, suggesting it sees data readiness as a key commercial opportunity if the new platform gains momentum.
The third law is aimed at speed, particularly for firms trying to test technologies in areas where the legal status is unclear. Under the amendment to the Special Act on Promotion of Research and Development Zones, technologies that are not explicitly prohibited by law can receive regulatory sandbox exemptions through review by a professional committee, without passing first through the Research and Development Zone Committee. That could cut the review process from roughly 90 to 150 days to around 30 to 60 days, making field testing faster for AI, robotics, autonomous driving and digital health projects.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
