grceng.academy
Policy lab 45 bonus XP

Gate high-risk model deployments

Start labRead the room

Objectives

An AI governance policy exists as a slide deck. Meanwhile a model that screens job applicants shipped without an evaluation, and another shipped with no human review of its decisions.

Turn the policy into a deployment gate: high-risk model releases are denied without a completed evaluation, without human oversight, or with unreviewed training data, while internal-use releases and fully governed high-risk ones proceed.

Release events the gate will see

#  use_case   eval_completed  human_oversight  data_reviewed  model_card
F1 high_risk  false           true             true           true
F2 high_risk  true            false            true           true
F3 high_risk  true            true             false          true
F4 high_risk  true            true             true           false
F5 internal   false           false            false          false

F1, F2 and F3 must be denied. F4 and F5 must proceed.
Fields arrive on input with exactly the names above.

Checks the server runs (5)

  • Rules parse as deny blocks
  • Missing evaluation is denied
  • Missing human oversight is denied
  • Unreviewed training data is denied
  • Governed and internal releases proceed

Check names are all you see here: the expectations that decide pass or fail, and the reference solution, stay on the server.

Step-by-step hints

The guided walkthrough, in full. It is free: no XP cost, and it does not touch the paid hint ladder or the reveal gate.

  1. Step 1

    Re-read the brief and name the trap

    A governance policy that lives in a deck cannot stop a release. The exercise is translating three of its clauses into conditions the release pipeline can evaluate.

    Note which clause is deliberately absent from the brief.

  2. Step 2

    Read the fixture as a truth table

    Compare each denied release to F4 and identify the single field that differs. That field is the body of one block.

    F5 flips every field to the unsafe value and must still ship, which tells you what condition every block has to carry.

    Rules parse as deny blocksGoverned and internal releases proceed

  3. Step 3

    Write one block per clause

    Three clauses, three blocks, each scoped to the risk tier that the policy actually governs.

    Give each block a message an engineer can act on: which governance artefact is missing, not that the policy failed.

    Missing evaluation is deniedMissing human oversight is deniedUnreviewed training data is denied

  4. Step 4

    Resist the fourth field

    A missing model card is a real gap and belongs to a documentation control with its own owner. Denying on it here makes this gate's message wrong and blocks a release the brief says must proceed.

    Governed and internal releases proceed

Rules summary

  • XP on pass

    45 bonus XP the first time every check goes green.

  • Paid hints

    2 available. Each costs 25% of the lab XP (about 11 XP).

  • Reveal solution

    Scores 0 XP and unlocks only after 3 check runs.

  • Attempt cap

    60 check runs per learner. The runner warns you in the last 10.

  • Drafts

    Your editor content is saved in this browser, so a reload will not lose work.

  • Walkthrough

    Always free and never penalised: use it before spending a hint.