Operators of high-impact AI must be able to explain outputs and the main criteria behind them

Practical action

Document the decision criteria and training-data overview for each high-impact system and prepare a user-facing explanation format.

Who does it apply to?

AI business operators providing high-impact AI or products and services based on it.

Actors

Provider / developerDeployer / user organisationPublic authority / government body

Sectors

Cross-sector / all sectorsEmployment and HRFinancial services and creditHealthcare and life sciencesEducation and trainingPublic services and government

Use cases

AI in hiring and employmentAI in healthcareAI in finance, credit and insuranceAI in educationAI in public services

Applies from: 22 January 2026

Which controls meet this duty?

Satisfies: the control, operated properly, does the work the duty asks for. Supports: it contributes but the duty needs more. Each control page lists every other duty it serves, so work done once can be counted once.

• satisfiesProcessCustomer operations lead · once per ai system
  Decision explanation, human review and appeal route

  Serves 12 recorded duties · evidence: Adverse-decision explanation template, AI decision challenge and human review procedure, Challenge and reversal log

  Explanation of results and criteria to affected people.

• supportsProcessProduct or model owner · at launch and on material change
  Technical documentation, model cards and instructions for use

  Serves 14 recorded duties · evidence: Technical documentation file, Model card or deployer information pack, Instructions for use

  Training-data overview and decision logic documented.

Templates that cover this duty

Generated from the records. Free · no account · links emailed to your work address. This duty is cited in each file with its source reference and a link back here.

• AI Agent Registry and Permission Matrix

  Register · XLSX and DOCX · v6

• EU AI Act Article 50 Transparency Kit

  Kit · DOCX and XLSX · v44

• ISO/IEC 42001 Gap Assessment and Statement of Applicability

  Assessment · XLSX · v3

• NIST AI RMF ↔ EU AI Act ↔ ISO/IEC 42001 Crosswalk

  Crosswalk · XLSX · v5

All 10 templates that cover it →

What evidence would a reviewer expect?

Framework mappings

Original editorial crosswalks. They cite clause numbers only and reproduce no standard text; confidence reflects how direct the mapping is.

See every South Korea duty mapped this way →

Frequently asked questions

Cite this record

AIPolicyTracker (2026). “Operators of high-impact AI must be able to explain outputs and the main criteria behind them (AI Basic Act)” [AIPT-OBL-0064, verified 26 September 2026 by Bhaskar Bhatt]. https://aipolicytracker.org/id/AIPT-OBL-0064 (accessed 8 November 2026). Data licensed CC BY 4.0.

Permanent id AIPT-OBL-0064: it never changes or gets reused, and its address keeps working if this page moves.

Cite the official text alongside it: Framework Act on the Development of Artificial Intelligence and Establishment of a Foundation for Trust (AI Basic Act, Act No. 20676), Korea Ministry of Government Legislation, National Law Information Center, https://www.law.go.kr/lsInfoP.do?lsiSeq=268543.

@misc{aipolicytracker2026_operatorsofhighimpactaim,
  author       = {{AIPolicyTracker}},
  title        = {{Operators of high-impact AI must be able to explain outputs and the main criteria behind them (AI Basic Act)}},
  year         = {2026},
  howpublished = {\url{https://aipolicytracker.org/obligations/south-korea-ai-basic-act-art-34-high-impact-ai-explanation-measures}},
  url          = {https://aipolicytracker.org/obligations/south-korea-ai-basic-act-art-34-high-impact-ai-explanation-measures},
  urldate      = {2026-11-08},
  note         = {Data licensed CC BY 4.0. Accessed 8 November 2026}
}

Change history

Similar obligations in other instruments

• Employers must inform workers and their representatives before using high-risk AI at work — EU AI Act, European Union
• Disclose AI interaction and label synthetic content — EU AI Act, European Union
• Deployers must tell natural persons that a high-risk AI system is used in decisions about them — EU AI Act, European Union
• Deployers of emotion recognition or biometric categorisation must inform exposed persons — EU AI Act, European Union
• Deployers must disclose deepfakes and AI-generated text published on matters of public interest — EU AI Act, European Union
• Providers of generative AI must mark synthetic output as artificially generated in a machine-readable way — EU AI Act, European Union
• Deployers must explain individual decisions taken with high-risk AI on request — EU AI Act, European Union
• Frontier developers must publish a transparency report before deploying a new frontier model — California SB 53, California (United States)