Rule set sha256:5f51c649da68

rego-art6-3-001

One rule of the rule set with digest 5f51c649da68af0ecaa5c68cf661a38494e679fd256373f2ef9f959f54671693. An audit record that names this digest and this rule was decided by the rule as shown here. Every rule in this set.

rules/rego/complyedge/article6/art6_3_procedural_derogation.rego

Citation
Regulation (EU) 2024/1689, Article 6(3)
Condition
An Annex III system is not high-risk where it performs a narrow procedural task, improves a previously completed human activity, detects decision patterns without replacing human assessment, or performs a preparatory task. Claiming the derogation while describing substantive automated decisioning (replacing human assessment on credit, hiring, benefits, etc.) is a misclassification risk.
Layer 1 detects
within one line and in this order, "not high-risk" (also "high risk"), "6(3)" or "narrow procedural", then replace (any ending) followed by "human", or automated (any ending) followed by credit, loan, hiring, recruit, welfare, benefit or insurance and then "decision"; claim, assert or invoke, then "6(3)", "narrow procedural task" or "preparatory task", then "without human" oversight, review or assessment, or "fully automat"; and derogat (any ending), then "annex iii", then "credit scor", creditworthiness, "hiring decision" or "insurance pricing". It matches wording, not meaning: it misses a misclassification told in other words, and it fires on a text that claims the derogation and says the system does not replace human assessment ("Our CV parser is not high-risk under Article 6(3): it only converts resumes into structured fields and does not replace human assessment; recruiters make every decision."). Whether the derogation is claimed for a system that does replace human assessment is judged in Layer 2.
Condition type
deterministic
Status in the file
approved