DR & AJU Legislation & Government Relations Consulting Center's latest issue of its specialized legislative journal, the Policy & Business Report, is now available in digital format.
Movements to regulate AI are expanding… Will there be progress in the legislative discussion of the framework act on AI pending in the National Assembly?
Expectations of the innovation that generative artificial intelligence (AI), such as ChatGPT, will bring continue to rise, while there are also growing voices urging to take precautions against potential risks. Major countries are accelerating their regulations on AI, with the European Union (EU) in the lead. Korea also announced its plan to legislate laws on AI regulation to present international standards. However, the National Assembly of Korea, which is in charge of finalizing legislation, is currently absorbed in the political strife between the ruling and opposition parties and keeps pushing back the legislation process. Some critics say that Korea needs to urgently establish a system to enhance its AI competitiveness while minimizing side effects amidst the fierce hegemonic competition in AI. Since each country has different standpoints and conditions, there is also a growing recognition of the need for cooperation among the public and private sectors worldwide under the principle of personal information protection.
The United States federal government has been pushing out laws, executive orders, and guidelines, which commonly aim at protecting user rights and managing potential risks of AI. Such an idea can be found in the “AI Risk Management Framework (AI RMF)” published this year by the National Institute of Standards and Technology (NIST) under the US Department of Commerce. The AI RMF provides guidelines for managing bias in AI. The recently published SAFE Innovation Framework for AI policy by Senator Chuck Schumer, named as an acronym for the key principles of AI regulation—security, accountability, foundations, and explainability— suggests the future direction for AI regulation.
The EU’s AI regulations are regarded as more severe than that of the US. On the 14th, the European Commission, the administration of the EU, approved a draft of a landmark bill for the first regulation enhancing transparency and safety of generative AI, such as ChatGPT, and began trilateral negotiations between the European Commission and the board of directors of 27 member countries. According to the draft, AI developers are required to disclose the sources of data used for AI model training and explicitly state the fact that the content was generated by AI, not humans. If the data sources are disclosed, those who hold copyrights of such data can demand revenue sharing or the usage fee of such data from AI development companies.
China is also swiftly pushing forward with legislation to regulate AI. The regulation stipulates that AI-generated content shall reflect the core value of socialism and shall not disrupt social order. In addition, it is also stipulated that AI-generated content is required to comply with the People’s Republic of China Network Security Law and Personal Information Protection Law. If such content fails to meet the requirements, the AI service will be suspended. This regulation will be finally introduced after going through the Standing Committee of the National People’s Congress of China.
<Regulatory Bills on AI in Major Countries>
Domestic AI regulation just made a start. The Act on the Fostering of Artificial Intelligence Industry and Trust Foundation Establishment, which integrates seven related bills proposed by members of the 21st National Assembly, is currently pending in the Science, ICT, Broadcasting, and Communications Committee of the National Assembly.
This bill, which includes opinions from related government departments, including the Ministry of Science and ICT, mainly aims to establish national strategies to support the advanced development of AI technology. This bill sets “pre-approval and post-regulation” as the fundamental principle for AI technology development and provides measures to ensure reliability by designating high-risk utilization areas that may harm public safety, such as automatic driving.
However, this bill has not been discussed after passing the Legislation and Judiciary Committee of the Science, ICT, Broadcasting, and Communications Committee of the National Assembly in February. While there is a consensus on the need to improve the legal system in response to the rapidly advancing AI technology between the ruling and opposition parties, the actual legislative discussions seem pushed aside. Some claim that lawmakers disregard matters that require legislation due to the prolonged confrontation over contentious bills and political issues between both parties.
Amidst this situation, the Personal Information Protection Commission (PIPC) and the Presidential Committee on the Digital Platform Government held an international conference titled “Artificial Intelligence and Data Privacy: Tackling Newly Emerging Challenges” on the 23rd. The conference was attended by personal data protection authorities of major countries (Korea, UK, Germany, and Japan), international organizations (EU and OECD), AI experts at home and abroad, and about 300 industrial, academic, and research officials and the general public.
In particular, this conference is highly regarded as significant in that the PIPC initiated discussions on the future direction of AI regulation with the personal data protection authorities of major European countries (UK, Germany, and France), who are leading discussions on the global regulatory systems in the AI field.
In order to address AI-related issues, the PIPC presented three key policy directions: (i) providing safety measures for AI service providers; (ii) shifting paradigm from rule-based regulation to principle-based regulation; and (iii) establishing differential regulations based on the risk of AI. The PIPC also announced its plans to establish communication channels between agencies for AI data privacy issues in various countries.
As the first-ever Korean law firm to issue specialized legislative journals, DR & AJU LLC has been publishing the monthly Policy & Business Report since August 2019. The May 2023 issue of the P&B Report conducted a full inspection of legislation proposed to the National Assembly Subcommittee, Standing Committee, and Plenary Session from May 15, 2023, to June 14, 2023, and selected and analyzed legislation with a significant impact on corporate activities. In the case of major legislation, key mentions of legislators, members of the Standing Committee, and government officials are included.
*DR & AJU Legislation & Government Relations Consulting Center provides P&B Report subscription service. Please contact us at Ic@draju.com for further inquiries.
