South Korea - AI Industry Promotion (AIPTRXX/2022)

Overview

The 'AI Industry Promotion and Trust‑Building' bill submitted by Representative Yoon Doo‑hyun on 7 December 2022 was framed as a dual‑track legislative proposal that combined statutory measures to promote South Korea's AI industry with baseline legal duties to secure public trust in AI products and services. The bill gained attention because it explicitly embodied a 'prior permissive, subsequent regulatory' approach (Korean: 우선허용·사후규제) — an orientation aimed at rapid industrial deployment while reserving ex‑post corrective powers for harms. That design made the bill a nucleus for subsequent legislative consolidation and debate. For a snapshot of the bill and its role in later legislative consolidation see the National Assembly information portal and contemporaneous analyses such as press coverage and stakeholder submissions. For the National Assembly bill information system see National Assembly — Bill Search. For stakeholder input on this specific proposal see the Business Software Alliance submission at BSA (submission on bill #18726).

Definitions

The proposal defined key terms to draw regulatory boundaries: 'Artificial Intelligence' (broad systems that perform tasks using algorithms and data), 'high‑risk AI' (AI used in areas with significant potential impacts on life, safety, fundamental rights or public order), 'AI business operator' (entities that develop, provide, or operate AI systems), and 'generative/large‑scale AI' (models that produce text, images or other content). The draft typically aligned definitions pragmatically with existing Korean IT and data laws to facilitate inter‑operation across ministries and avoid duplicative definitions. Definitional choices were central to later debate because they determine the suite of systems subject to mandatory safeguards and disclosure duties.

Governance and Institutional Framework

The bill proposed a layered governance architecture: (1) a national coordinating mechanism to set strategic priorities and the multi‑year basic plan; (2) a national AI support center to execute promotion policy (grants, tech assistance, clusters); and (3) delegated authority to line ministries for sectoral enforcement. In committee debates these roles were debated intensively with proposals to place the oversight body under the President, the Prime Minister, or as an expert council. The practical duties assigned to the Ministry of Science and ICT (and related agencies) included preparation of a 3‑year AI basic plan, standards and conformity coordination, and operational oversight of reliability testing and certification programs. See contemporaneous reporting on committee consolidation for details: Maeil Business (MK) and legislative trackers such as Hankyung — Legislative Explorer.

Key Focus Areas

The proposal focused on the following substantive policy areas: industry promotion (funding, cluster/center supports, public‑private partnership incentives), tiers of AI risk governance (lighter rules for low‑risk experimentation; higher obligations for high‑risk uses), transparency (notice to users when services rely on AI and explicit labelling for generative AI outputs), trust and reliability (documentation, technical records, risk management systems), conformity and certification (voluntary-to-mandatory pathways for high‑risk AI), and enforcement (administrative corrective powers and sanctions for non‑compliance). The bill's high‑risk category list (healthcare, finance, employment, public administration, critical infrastructure) reflected international practice and was a central point of debate because it triggered stronger duties for operators. For press analysis of the tension between innovation and safety see AJU News.

Implementation Framework

Implementation was designed to be operationalized by the Ministry of Science and ICT in cooperation with other ministries and a designated national AI center. The bill included provisions for: issuing implementing regulations (presidential/delegated acts) to specify technical standards for disclosure and documentation; a phased approach for sectoral implementation of high‑risk obligations; pilot certification programs to foster domestic conformity assessment capacity; and funding authorizations for industry promotion and public interest research. The framework envisioned a mix of obligatory duties for clearly‑defined high‑risk applications and supportive measures (grants, training, R&D tax incentives) to accelerate domestic capability building.

Monitoring and Evaluation

Monitoring relied on periodic reporting obligations by AI business operators, obligations to retain technical documentation and logs, and government fact‑finding powers to request or inspect records. The bill proposed 3‑year strategic reviews tied to the national AI basic plan and required publication of progress reports to improve policy transparency. It also envisioned metrics for industry growth and safety incidents to evaluate whether the preferential promotional regime was generating acceptable trade‑offs between innovation and public protection.

Penalties, Liability, and Appeals

To balance promotion with enforceability the bill established administrative enforcement tools: corrective orders, administrative fines, temporary suspension of services and, in serious cases, referral for criminal investigation under existing statutes where willful misconduct or gross negligence caused harm. It further provided for administrative review and judicial appeal rights for regulated entities. The exact quantum of fines and procedural details were left to implementing regulations or later legislative consolidation, which became a point of negotiation in committee proceedings.

Relationship to Other Instruments

The bill was drafted to interact with South Korea's existing legal ecosystem: the Personal Information Protection Act (privacy/data), telecom and electronic communications law (network safety), consumer protection statutes, product safety frameworks, and sectoral rules (health, finance). It proposed regulatory coordination mechanisms to reduce overlap and set rules for cross‑border data flows when relevant. The need to align with other domestic laws and with international instruments (e.g., EU AI Act trends) was frequently highlighted by stakeholders.

International Alignment

The bill explicitly sought to position South Korea competitively with major AI jurisdictions by adopting a pro‑innovation stance while importing elements of international best practice (risk‑based focus, transparency duties for generative systems, conformity pathways). Internationally‑oriented provisions included support for cross‑border testing programs, recognition of foreign certificates where appropriate, and encouragement for industry to comply with international standards to facilitate export competitiveness. Commentators noted parallels and divergences with the EU AI Act and UK 'pro‑innovation' approaches, and recommended continued alignment where practicable.

Implementation Timeline

Sources and References