South Korea officially rolled out the AI Basic Act, marking what the government describes as the world’s first comprehensive legal framework for artificial intelligence (AI). The move, however, has reportedly sparked immediate pushback from local startups that fear “regulatory risk” could stifle innovation before it even begins. According to a report by news agency Reuters, through this act, the country aims to secure its place as a top three global AI powerhouse.This also means that the launch also puts South Korea ahead of the European Union (EU), whose landmark AI Act is being implemented in slower phases through 2027. Seoul’s unified law consolidates 19 separate regulatory proposals and is now fully operational across both the public and private sectors.
What does the new rules say
Under the South Korea’s AI laws, companies must ensure human oversight in “high-impact” AI areas such as nuclear safety, the production of drinking water, transport, healthcare and financial uses such as credit evaluation and loan screening.Another rule says that AI companies are required to provide users advance notice about products and provide clear labelling or watermarking for AI-generated output, which is difficult to distinguish from reality.
South Korean startups are not happy: ‘Why do we have to be first?’
The rules also provide minimum one-year grace period during which no administrative fines will be issued but the startup community is sounding the alarm. According to a survey by the Startup Alliance, only 2% of AI-focused startups feel they have a formal compliance plan in place. Moreover, about half of them admit they do not fully understand the new law.“There’s a bit of resentment — why do we have to be the first to do this?” Lim Jung-wook, co-head of the Startup Alliance, was quoted as saying. Founders are concerned about the law’s “vague” language, fearing it may force them to adopt overly cautious development strategies.President Lee Jae-myung, meanwhile, expressed sympathy for these concerns, urging policymakers to maximise the industry’s potential through institutional support while “pre-emptively managing anticipated side effects.”The Ministry of Science and ICT (MSIT) has launched a dedicated AI Act Support Desk to help companies determine their regulatory status. If companies fail to comply after the grace period, they may be fined up to 30 million won ($20,400), which the government notes is significantly more “moderate” as compared to penalties in the EU’s framework.
