The landscape of artificial intelligence (AI) in South Korea is facing significant challenges, particularly due to the soaring costs associated with acquiring quality learning data. Recent revelations indicate that domestic startups are now required to pay as much as 1 million won for a single book, which is often used as a primary source of learning data. This escalating price has made it increasingly difficult for small and medium-sized enterprises (SMEs) to access high-quality datasets, thus widening the performance gap between agile startups and their larger corporate counterparts.
As the market for usable data becomes more scarce, companies are encountering formidable barriers in their quest to develop competitive AI models. For instance, a notable AI startup that was in the process of developing a large-scale language model had to abandon its plans for utilizing scientific literature as training data due to exorbitant licensing fees. The legal framework surrounding copyright is particularly stringent in South Korea, lacking any exceptions for text and data mining, which forces organizations to secure separate licenses for every piece of content they intend to use for AI training.
In tandem with these financial pressures, the evolving landscape of privacy regulations is further complicating the situation. The recent amendments to the Personal Information Protection Act have made it more challenging for companies to leverage potentially valuable data sources, such as CCTV footage for analysis. Startups that had innovative AI business models based on video data analysis are now facing increased uncertainty due to the ambiguous risks surrounding re-identification penalties.
The government, acknowledging these hurdles, has expressed its commitment to facilitating the innovation of SMEs through targeted policy initiatives. Jo Joo-hyun, the director of the Small and Medium Business Administration, highlighted the importance of diversifying export markets and supporting overseas investments during a recent press conference celebrating his first year in office. He revealed that the budget for the upcoming fiscal year has been increased by 10.5%, totaling approximately 16.8 trillion won, with a significant portion allocated to support R&D and AI ventures.
However, despite these efforts, the fundamental issues surrounding data access and legal uncertainty remain unresolved. The Ministry of Economy and Finance has taken steps to alleviate some of the burdens in the semiconductor sector by proposing to reduce tariffs on key materials used for wafer production to zero percent. This decision aims to enhance the competitiveness of domestic semiconductor manufacturers in light of high tariffs imposed by the United States.
As we witness these developments unfold, it becomes clear that the combination of rising data costs and stringent regulatory frameworks poses a substantial challenge to the growth of the AI sector in South Korea. Businesses must navigate these complexities while simultaneously seeking innovative solutions to sustain their operations.
In conclusion, as the demand for AI learning data continues to surge, the need for a robust supply of quality public data becomes increasingly urgent. Policymakers must address the legal barriers and financial challenges that hinder access to data. Only through a concerted effort to enhance data accessibility and minimize legal risks can South Korea’s startup ecosystem thrive in the era of AI innovation.
[관련기사] https://n.news.naver.com/mnews/article/011/0004528420?sid=101
Leave a Reply