Skip Navigation
Menu
Newsletters

Legislative Trends in AI: National AI Ethical Standards and Amendment to Enforcement Decree of Framework Act on Intelligent Informatization

2020.12.23

On November 27, 2020, the Ministry of Science and ICT (the “MSIT”), with Korea Information Society Development Institute (KISDI), prepared a draft of the “National AI Ethical Standards” (the “Proposed AI Ethical Standards”), a set of broadly applicable general principles for presenting the direction for the development and utilization of AI.  On December 8, 2020, comprehensive amendment to the Enforcement Decree of the Framework Act on National Informatization (the “Enforcement Decree”) was promulgated as a follow-up measure to the wholly amended Framework Act on National Informatization (amended on June 9, 2020, the “Amended Act”). 

The Korean government had announced its “National AI Strategies” plan, which was created jointly by all government departments, including the MSIT, on December 17, 2019.  To achieve one of the goals of the National AI Strategies, namely, “realizing human-centered AI,” the MSIT drafted the Proposed AI Ethical Standards which define basic ethical standards that all members of society, including developers, suppliers and users, can refer to in all AI-related processes from development to use.  The Proposed AI Ethical Standards were created to be non-binding “ethical standards” to foster an environment of voluntary regulation in the relevant industries, and are expected to function as an AI ethics platform that allows for continuous updates and revisions of the ethical standards by reflecting AI ethics-related issues that may newly emerge with technological developments.

Meanwhile, through the Amended Act, the MSIT revised the title of the Framework Act on National Informatization to “Framework Act on Intelligent Informatization” based on its decision that the existing regulation must be completely revised to preemptively respond to societal and economic changes following the 4th Industrial Revolution, establishing a foundation for a system to facilitate the use of intelligent information technology.  Accordingly, the MSIT drafted the Enforcement Decree to define the specifics of matters stemming from the Amended Act, such as establishing and implementing policies for building an intelligent information society and supporting the generation, distribution and utilization of data, and necessary matters for their enforcement.

Below are the key details of the Proposed AI Ethical Standards and the Enforcement Decree:

1.    Key Details of Proposed AI Ethical Standards
 

  • Humanity as the principal value pursued by the Proposed AI Ethical Standards
     
    • All AI must pursue humanity as its principal value in support of “AI for Humanity,” meaning that AI must be developed and controlled to not only benefit humans but also preserve and cultivate the unique characteristics of humans without causing any damage.
       
  • The three basic principles of AI
     
    • As the basic principles to be considered throughout the entire process from the development and utilization of AI for AI to realize the traits of humanity, AI should be used in accordance with the principles of “protecting human dignity,” “achieving the common good for society” and “fulfilling the intended purpose of technology.”  
       
  • Ten core requirements of AI
     
    • Ten specific requirements for implementation must be observed in order to realize the aforementioned three basic principles of AI.
    • For example, the “principle of data control” states that (i) data that includes personal information must be used in a way that conforms to the corresponding purpose of the data and must not be used for any other purposes, and (ii) data quality and risk must be controlled to minimize data bias in all stages of data collection and use.


2.    Key Details of Enforcement Decree
 

  • The Enforcement Decree prescribes the scope of intelligent information technologies that must comply with technological standards set forth by the Minister of the MSIT that are closely related to the life or physical safety of the public (Article 16).
     
    • The applicable technologies would be “intelligent information technology that may pose a significant threat to the life and health of the user which (i) is intended to be developed, managed or utilized for military purposes, (ii) can affect a person’s body by being directly used in medical services such as surgeries and other procedures defined under Article 24-2 (1) of the Medical Service Act or (iii) can pose a serious threat to humans if it malfunctions.”
       
  • The Enforcement Decree stipulates procedures for determining standards on the performance and reliability of the information protection systems announced by the Minister of the MSIT, as well as matters regarding recommendations for improvement where the above standards are not met (Article 51).
     
    • When establishing and publicly notifying the standards for performance and reliability of the information protection system, the Minister of the MSIT must confer with the head of the relevant institutions in advance. 
    • In addition, if any entity that manufactures or imports the information protection system requests confirmation of whether the system conforms to the above standards, the Minister of the MSIT may request relevant investigation, or tests and inspections to the head of Korea Internet & Security Agency (KISA) or any institution meeting the standards prescribed by applicable international agreements.


Implications

The Proposed AI Ethical Standards of the MSIT were created to be ethical standards, not binding or enforceable regulations.  However, it is highly likely that they will come to serve as the point of reference for voluntary regulation standards applicable throughout the whole process of AI development and use.  The Proposed AI Ethical Standards are expected to be finalized in December at the latest after reviewing comments received until December 15 from the general public.  Considering that the Proposed AI Ethical Standards will likely be used as basic policy guidelines for AI-related government regulations, it is necessary to continue monitoring relevant legislative developments.

In addition, the Enforcement Decree took effect together on the effective date of the Amended Act, December 10, 2020.  Therefore, we advise closely following legislative trends in the relevant authorities’ amendment process while monitoring the legislation of subordinate regulations delegated from the Enforcement Decree, such as MSIT notifications and guidelines.

Share

Close

Professionals

CLose

Professionals

CLose