Informational Fresh jira_story: GOV-115

Lesson 6 of 11 — Confirm or override EU AI Act classification

Governance Onboarding · Source: Jira Story GOV-115 (5 pts) · Authority: Informational (the Jira story is authoritative).

Purpose

Confirm the AI-suggested EU AI Act risk classification, or override it with a mandatory, attributed reason.

Why it matters

Risk classification drives which obligations apply; overrides must be defensible.

Where this lives

Runs in the Customer AI Governance platform (components: onboarding, intelligence). Classification is suggested by the intelligence layer; an override captures a mandatory reason plus actor and timestamp. Learn teaches this read-only.

Workflow

Given an AI classification suggestion
When it is overridden
Then the reason is mandatory and the actor/time are recorded; all in-scope systems must be classified to continue

Worked example

An assessor disagrees with the AI's 'high-risk' suggestion and sets 'limited-risk'. What must the system capture before continuing?

A mandatory reason plus who changed it and when — and all in-scope systems must be classified before the step completes.

Acceptance criteria (authoritative — GOV-115)

Given an AI classification suggestion, when it is overridden, then the reason is mandatory and the actor/time are recorded; all in-scope systems must be classified to continue.

Common pitfalls

Overriding the AI suggestion without a recorded reason, breaking the audit trail.

Knowledge check

Scenario: An assessor disagrees with the AI's 'high-risk' suggestion and sets 'limited-risk'. What must the system capture before continuing?
Look for: A mandatory reason plus who changed it and when — and all in-scope systems must be classified before the step completes.

Key concepts

EU AI Act risk tiers
mandatory override reason
actor / time attribution

For the AI Tutor

Summary: Confirm the AI-suggested EU AI Act risk classification, or override it with a mandatory, attributed reason.
When to use: teaching or answering questions about "confirm or override eu ai act classification" during governance onboarding.
When NOT to use: choosing frameworks (classification feeds framework applicability, GOV-116).
Key concepts: EU AI Act risk tiers, mandatory override reason, actor / time attribution.

Completion criteria

Learner passes the scenario knowledge check and can point to the system state that satisfies GOV-115's acceptance criterion.

Module: LEARN-MOD-001 (Governance Onboarding)
Provenance: LEARN-MAP-001 · Generation standard: LEARN-STD-002
Next: LEARN-LSN-007