REEU AI Act · Compliance
REGULATION(EU) 2024/1689
ENFORCEMENT2 Aug 2026
COVERAGEArt. 12 · 13 · 14 · Annex III

The architecture is the audit trail.
The Pack is the encoded specifics.

Ctrl AI doesn’t answer your question and then write a log file about it. The answer is a sequence of signed Ctrl invocations, each producing an Ed25519 receipt anchored in an append-only transparency log. Article 12 record-keeping, Article 13 transparency, and Article 14 human oversight come for free with the architecture. The EU AI Act Starter Pack adds the typed Ctrls that encode the Act’s specific decisions — risk-tier classification, oversight gating, retention checks.

Included on Starter and above·Signed by your compliance officer on install·Named external co-signing on Business and Enterprise
§1 · Timeline

The enforcement clock is running.

  • 2024-08-01Regulation entered into force
  • 2025-02-02Prohibited-AI ban + AI literacy obligation
  • 2025-08-02GPAI model obligations apply
  • 2026-08-02Most obligations enforceable · fines begin
  • 2027-08-02Existing high-risk systems must comply

Fines up to €35M or 7% of global annual turnover

§2 · Requirements

What the regulation actually asks for.

Art. 12

Record-keeping

High-risk AI systems must automatically record events for traceability. Records must be appropriate for the system's intended purpose; deletable only after at least six months.

Art. 13

Transparency to deployers

Systems must be designed for sufficient transparency so deployers can interpret outputs. Instructions for use, characteristics of training data, expected accuracy, and human oversight measures must be documented.

Art. 14

Human oversight

Effective oversight by natural persons. The system must allow the human to override, intervene, or interrupt. Automated decisions affecting natural persons require a verified path to human review.

§3 · Solution

What the architecture gives you. What the Pack adds on top.

The Act’s structural obligations — record-keeping, transparency, human oversight, robustness — are satisfied by how every Ctrl runs in Ctrl AI. You don’t install them; they are the platform. The Pack is the layer above: typed Ctrls that encode the Act’s specific decisions, so when your team asks an AI question, the agent has signed functions to call instead of opinions to render.

What the architecture satisfies — for every customer, no install required
  • Art. 12

    Automatic record-keeping

    Every Ctrl invocation writes a receipt row and a transparency-log entry. The log is hash-chained — past entries cannot be rewritten without breaking every entry that follows. Six-month retention is a config minimum, not a feature you have to remember to turn on.

  • Art. 13

    Sufficient transparency to deployers

    A Ctrl is a typed function — explicit inputSchema, outputSchema, post-conditions, human-signed rationale. The deployer reads exactly what will run. The LLM does not paraphrase prose into behaviour; it picks a Ctrl and calls it.

  • Art. 14

    Effective human oversight

    Destructive Ctrls refuse to invoke without an approved gate. The dispatcher enforces this — the model cannot bypass it. Delegations from a human to an agent or external API client are scoped by Ctrl, jurisdiction, condition, and expiry.

  • Art. 15

    Accuracy and robustness

    Post-conditions run after every invocation; a failure degrades the receipt's trust level (verified → partial → degraded). Typed inputs and outputs are validated at the boundary — malformed calls are rejected before they reach the runtime.

What the Starter Pack adds — typed Ctrls for the Act’s specific decisions
  • classify-ai-use-case-risk-tier — Annex III risk-tier decision (unacceptable / high / limited / minimal)
  • human-oversight-required — Article 14 gate decision keyed off risk tier
  • log-ai-action-decision — Article 12 record-write with retention metadata
  • retention-policy-check — Article 12 check against your org's policy days
§4 · Walkthrough

A real question, traced end to end.

Below is what happens when a compliance officer at a mid-sized European bank (the fictional acme-bank) asks Ctrl AI a real question. Each step is a Ctrl invocation; each invocation produces a receipt; each receipt links back to the EU AI Act clause it satisfies. Nothing is paraphrased.

/ask · acme-bank · 2026-05-13 09:42 UTC
Question

“Can we deploy this credit-scoring model to score retail mortgages for our French branch?”

01
search_ctrls· ranking

Agent searches the installed catalog with “credit scoring EU AI Act risk classification”. Because acme-bank installed eu-ai-act-starter@v1 and the org is jurisdiction-tagged [“EU”], three Ctrls rank highest, sorted by clause match.

classify-ai-use-case-risk-tier   [appliesToJurisdictions: ['EU']]
human-oversight-required          [satisfiesClauses: article_14]
log-ai-action-decision            [satisfiesClauses: article_12]
02
invoke_ctrl· grounded_llm

Agent calls classify-ai-use-case-risk-tier with the use-case description and French retail context. The grounded-LLM runtime cites Annex III as a source and returns JSON-schema-validated output. Two post-conditions pass.

{ riskTier: 'high',
  reasoning: 'Annex III §5(b) covers AI used to evaluate
              creditworthiness of natural persons.',
  citedAnnexClauses: ['Annex III §5(b)'] }
trust = grounded · receipt rcp_a91f… signed · log seq #18,402
03
invoke_ctrl· deterministic

Agent chains the result into human-oversight-required using {{steps.02.outputs.riskTier}}. The deterministic runtime returns a typed boolean — no model in the loop.

{ oversightRequired: true,
  rationale: 'Article 14 mandates meaningful human oversight
              for this risk tier.',
  citedArticles: ['eu_ai_act:article_14'] }
trust = verified · receipt rcp_b73c… signed · log seq #18,403
04
invoke_ctrl· deterministic

Agent records the decision via log-ai-action-decision so the determination itself is recorded under Article 12 — not just the question that triggered it. The record id becomes the audit anchor.

{ recordId: 'rec_8h2k4nq1',
  loggedAt: '2026-05-13T09:42:14.221Z',
  satisfies: ['eu_ai_act:article_12',
              'eu_ai_act:article_13'] }
trust = verified · receipt rcp_c04e… signed · log seq #18,404
Answer rendered to the user

This deployment is classified high-risk under EU AI Act Annex III §5(b) [ctrl:classify-ai-use-case-risk-tier#a91f]. Article 14 therefore requires meaningful human oversight before deployment [ctrl:human-oversight-required#b73c]. The determination has been recorded under Article 12 as record rec_8h2k4nq1 [ctrl:log-ai-action-decision#c04e]. You need an approved oversight gate from the DPO before this can ship.

Every [ctrl:slug#id] chip is clickable. Each opens the receipt, shows the Ed25519 signature, and links to the transparency-log entry on the public Trust Portal. None of the factual claims above can exist without a signed receipt to back them.

What the regulator sees if they ask: an append-only chain, four receipts, each signed, each tied to the article it satisfies, each independently verifiable against a publicly published public key. No screenshot evidence. No after-the-fact log reconstruction. The audit artifact is produced by the system, not collected from it.

§5 · Pack contents

The Starter Pack — four typed Ctrls.

The v1 Starter Pack ships the four highest-leverage decisions for early deployers. Each Ctrl has a typed input schema, a typed output schema, executable post-conditions, and a cited regulatory clause. Install materialises them into your org’s library, where your compliance officer refines and signs each one. The pack grows over time — v2 adds Annex IV technical documentation, conformity-assessment workflow, and GPAI obligations.

SlugNameClauseRuntime
classify-ai-use-case-risk-tierClassify AI Use-Case (Risk Tier)Annex IIIgrounded_llm
human-oversight-requiredDecide if Human Oversight is RequiredArt. 14deterministic
log-ai-action-decisionLog AI Action / DecisionArt. 12 · 13deterministic
retention-policy-checkRetention Policy Compliance CheckArt. 12deterministic
Authored by
Ctrl AI compliance team
Signed by
Your DPO / Head of Risk, at install
Review cadence
180 days · auto-flag drift

Five minutes to install.
Eighteen months to ignore.

The clock is the regulator’s, not ours. Install the Pack today; customise the Ctrls to your context; ship signed AI before Aug 2026.