The importance of the AI Basic Act

Home > Opinion > Columns

print dictionary print

The importance of the AI Basic Act

Lee Kwang-hyung
 
The author is president of KAIST and cochair of the Presidential Council on Intellectual Property.
  
On Dec. 26, last year, the "Basic Act on Artificial Intelligence (AI) Development and Trust Building" passed the National Assembly with overwhelming support: 260 votes in favor, one against and three abstentions out of 264 attendees. Often referred to as the "AI Basic Act," this legislation will take effect one year after its proclamation, with detailed provisions to be established through enforcement ordinances. Korea's enactment of AI-related legislation follows the European Union (EU), making it the second country globally to do so.
 
It is almost surprising to see such significant progress from the National Assembly, often perceived as a place of constant partisan conflict. Particularly noteworthy is the work of the Science, ICT, Broadcasting and Communications Committee, which, despite the turbulence of martial law and impeachment crises, collaborated to draft and pass a forward-looking bill. Their bipartisan effort for Korea’s future deserves commendation and sets an exemplary model for political cooperation.
 
Balancing regulation and promotion: A dual approach
Since the advent of ChatGPT in late 2022, AI has been rapidly reshaping society. At the recent CES in the United States, AI dominated discussions. Once focused on learning and reasoning, AI has advanced beyond language translation to creative and emotional processing. Next year, AI is expected to mimic human creativity and emotions, expanding its reach to cultural and defense domains.
 
As AI capabilities grow, nations and societies must define how to manage and utilize these transformative technologies. Korea’s legal framework historically made it difficult to respond to such changes, as new technologies often require explicit legal authorization for action. In this context, the AI Basic Act is a welcome development.
 
The United States and China currently have a laissez-faire approach to AI and have yet to establish AI-specific legislation. This appears to be rooted in confidence as both countries are global front-runners in AI development. Conversely, the EU’s regulatory-heavy AI legislation reflects its desire to avoid a repeat of its digital revolution setbacks. Korea’s AI Basic Act, in contrast, strives to balance regulation with promotion — a carefully considered approach.
 
Key provisions of Korea’s AI Basic Act
The AI Basic Act is significant for clearly outlining the direction for the AI industry's development and regulation. It defines the government’s role in fostering responsible AI advancement while preventing harm to society. The Act is structured around three key pillars.
 
First, the Act establishes the "National AI Committee" under the Office of the President, to be chaired by the president. The Minister of Science and ICT must develop and implement a triennial "AI Master Plan" outlining policies, work force development and AI initiatives. The committee will deliberate and decide on key matters related to AI promotion, infrastructure and strategic direction.
 
Second, to facilitate regulation and promotion, the Act defines legal terms such as "high-impact AI" and "generative AI." High-impact AI refers to AI systems with significant implications for human life, safety and essential services, such as energy supply, health care, transportation and public services. Generative AI pertains to systems like large language models (LLMs) that create outputs such as text, images and sounds based on input data.
 
Third, the Act mandates transparency by requiring businesses to inform consumers when products or services involve high-impact or generative AI. It also requires safety evaluations for systems exceeding a certain computational threshold, measures to manage potential risks and assessments of the impact on human rights.
 
Prioritizing job creation over excessive regulation 
Compared to the EU’s AI Act, Korea’s AI Basic Act emphasizes flexible regulation. The EU categorizes AI into four levels of risk — “unacceptable risk,” “high risk,” “limited risk” and “minimal risk” — and implements detailed controls. In contrast, Korea adopts a value-neutral approach by focusing on "impact" rather than "risk," regulating only AI with significant effects on human life and safety.
 
However, Korea’s scope of regulation is narrower than the EU’s. The Act exempts AI used solely for national defense or security purposes. By comparison, the EU excludes AI used exclusively for research, development or scientific purposes from its regulations. This difference could limit Korea’s ability to foster foundational research and innovation, as experimentation often requires unbounded creativity.
 
The AI Basic Act primarily sets the framework for promoting and regulating the AI industry. The specifics, including the criteria for defining high-impact AI and corresponding regulations, will be determined through enforcement ordinances. The success of Korea’s AI industry now hinges on these ordinances.
 
To this end, the government plans to establish a task force (TF) comprising experts from academia, industry, research institutions and government agencies. The TF will focus on defining and regulating high-impact AI as emphasized in the Act.
 
One critical caution is to avoid mirroring the EU’s defensive regulatory stance. Overregulation could stifle industry growth, curbing job creation and harming the public. Striking a balance between mitigating risks and fostering AI-driven job creation is crucial.
 
Given the interconnected nature of the global market, Korea must consider the regulatory environments of competing nations. Industries naturally gravitate toward countries offering more favorable conditions. Thus, regulations must be carefully calibrated to support domestic industries while ensuring global competitiveness. In preparing the enforcement ordinances, I urge policymakers to adopt a "pro-AI" and "pro-jobs" perspective. The ultimate goal should be to harness AI’s potential to enhance economic opportunities while minimizing its risks. 
 
Translated using generative AI and edited by Korea JoongAng Daily staff.   
 
 
Log in to Twitter or Facebook account to connect
with the Korea JoongAng Daily
help-image Social comment?
s
lock icon

To write comments, please log in to one of the accounts.

Standards Board Policy (0/250자)