§ Finance · AI Basic Act
Finance: AI Basic Act compliance & Korean LLM choice
When LLMs touch high-impact tasks like lending/underwriting, the documented rationale for your model choice becomes a core artifact for impact assessment and audit.
Why this sector is high-impact
Finance is treated as high-impact (lending, underwriting, scoring), and regulators are tightening AI guidance. As an adopter you carry duties, so model-selection rationale is a regulatory artifact.
Key duties
- ✓Impact assessment for high-impact AI
- ✓Safety/decision record retention
- ✓Transparency & explainability
- ✓Data sovereignty (network separation, sub-processing)
✦ Model-choice angle
Network-separation and data-sovereignty needs can favor self-hosting over direct global APIs. Prioritize commercial open weights (A.X 4.0) or EXAONE (license-checked) on-prem; global APIs need sub-processing/cross-border review.
Related
References
General information, not legal advice. Verify applicability and duties with official MSIT guidance and legal/compliance counsel.