“The key to competitive advantage in artificial intelligence (AI) lies in securing data. For corporations to advance AI technology, swift regulatory innovation regarding data use is needed.”
On the 25th, at a field meeting titled ‘Innovation in data regulation for becoming a top three AI nation’ held at the SW Maestro Training Center in Mapo, Seoul, chief executive officers (CEOs) from corporations in the AI and data sectors, including Upstage, Select Star, Twelve Labs, DOUZONE, and MAUM.AI, expressed a united voice.
They pointed out that “Various data utilization regulations continue to be a constraint for startups that hold innovative ideas in securing data, thereby hindering their innovative activities.” They emphasized that “It is necessary to create paths through regulatory sandboxes to freely utilize highly reliable data held by public institutions and public enterprises, original data needed to dramatically improve AI model performance, and public works.”
Kim Sung-hoon, CEO of Upstage, a developer of the large language model ‘Solar,’ said, “I hope the government creates a data pool and initially allows us to use data freely while keeping a record of what data is used, so that later, when revenue is generated, we can be charged based on how much data we used.”
Song Ho-cheol, head of the platform division at DOUZONE, an enterprise resource planning (ERP) provider, said, “Startups need to access data that is profitable, but there are many ambiguous aspects in current regulations. Even though there is no clause saying not to use the data, it often isn’t clear if using it is permissible.” He added, “Digitizing produced data is necessary for becoming an AI powerhouse, but it is also important for corporations to effectively consume the data that has already been established.”
Jung Jin-woo, co-founder of Twelve Labs, a video understanding AI development company, said, “It is not easy to distinguish what is personal data and what is not from video information,” and added, “I hope a regulatory data sandbox can be created to allow the use of de-identified personal data for securing data.”
Kim Se-yeop, CEO of Select Star, a data processing and construction company, said, “Due to the increasing importance of data, copyright holders want their value to be maximally recognized, but corporations have limited budgets, leading to a widening gap in perspectives.” He lamented that “For AI development corporations, difficulties arise because it has not been determined whether revenue will be generated when their business model is commercialized. This is also a challenging issue for companies that distribute data.”
Choi Hong-seob, CEO of MAUM.AI, a physical AI development company, said, “It becomes problematic when someone posts copyrighted data claiming it is their own, but it is not problematic to learn from and imitate it. The issue arises when it is used as ‘my work’ that I created.” He added, “I don’t understand why I have to pay to learn from data. If the data is public and viewable by everyone, regulations should be relaxed to allow learning from this data, but it would be better if regulations are imposed at the usage stage.”
Korea Startup Forum representative, attorney Koo Tae-eon, noted, “Due to regulations like the Personal Information Protection Act and Copyright Act that obstruct data distribution, there are no corporations capable of developing foundation-level large language models except for Naver,” and added, “In contrast, China has secured world-class facial recognition AI technology through the supply of facial information from 1.1 billion people.”
Ryu Je-myung, the second vice minister of the Ministry of Science and ICT, who presided over the meeting, stated, “The competitiveness of AI depends on how easily large amounts of high-quality data can be secured and utilized, and the new government is preparing to promote innovation in data regulation as a core national agenda.” He added, “I believe that regulatory innovation in data is an urgent task for AI advancement.”