Article 4
Context-specific literacy
Measures should reflect staff knowledge, use context and affected people. A generic badge is not the operating goal.
AI literacy evidence for EU agencies
Turn everyday AI use into an owned, reviewable practice.
DiscloseReady maps your agency’s real AI use, builds role-specific literacy material, records proportionate measures, and screens five external-facing cases without pretending every AI-assisted output needs a label.
The obligation is real. The evidence should be proportionate.
Article 4 AI-literacy measures have applied since February 2025. Official Commission guidance says staff using ChatGPT for advertising copy or translation should understand specific risks—and that an internal record of training or other initiatives can be retained. General AI Act application begins 2 August 2026.
Article 50
Specific transparency cases
Direct interaction, synthetic media and selected public-interest text require different facts. Provider and deployer roles matter.
Our boundary
No universal AI label
We screen actual use cases, record ambiguity and hold legal conclusions for qualified professional review.
One bounded sprint. Editable operational evidence.
The pilot starts after complete intake and cleared payment. You receive source files—not a black-box score or a claim that your organization is compliant.
€750
Fixed pilot fee. Tax treatment is shown on the invoice. 100% before kickoff.
- 01Register and normalize up to 12 AI use cases.
- 02Build role-based literacy material for up to 3 role groups.
- 03Supply a 12-question knowledge check, initiative log and completion-record template.
- 04Conservatively screen up to 5 external-facing AI use cases.
- 05Write disclosure options only where a plausible action is identified.
- 06Deliver prioritized implementation tickets with owners, evidence and acceptance criteria.
From messy use cases to an owned routine.
The workflow is intentionally small enough to complete in five business days and explicit enough for another reviewer to inspect.
Data-minimized intake
You describe purpose, roles, inputs, outputs, human review, external exposure and ownership. No passwords, private prompts or customer datasets.
Role and risk mapping
Actual use cases become learning objectives for creative, account and leadership roles—not a generic lecture detached from work.
Conservative screen
Each selected external case is marked action, no action identified, or professional review required, with the source and assumptions recorded.
Evidence and tickets
Training records, exact tasks, owners and objective acceptance criteria make the work operable after delivery.
Fictional demonstration
See the reasoning, not a fake testimonial.
Example Agency A is invented. Its five use cases demonstrate why a support chatbot and synthetic presenter may need action while human-reviewed ad copy does not receive an automatic AI label.
Open the complete sample pack| Use case | Initial screen | Next action |
|---|---|---|
| Ad-copy assistant | No action identified | Retain editor approval evidence |
| Support chatbot | Action | Verify first-interaction notice |
| Synthetic presenter | Action | Place appropriate disclosure |
| Translation assistant | No action identified | Retain competent review |
What it is
- A proportionate evidence and implementation sprint.
- Editable registers, training material and task records.
- A written clarification round after delivery.
- A practical starting point for agency ownership.
What it is not
- No legal advice, certification or compliance guarantee.
- No high-risk-system or conformity assessment.
- No technical provenance implementation.
- No changes inside your CMS, chatbot or ad accounts.
Questions worth asking before buying.
The fixed scope is designed to make “not a fit” as clear as “ready to start.”
Do we receive a compliance certificate?
No. The completion record documents participation in an internal initiative only. The Commission says Article 4 does not require a specific certificate, while organizations may retain internal records of training or other guidance.
Will every AI-assisted asset receive a disclosure label?
No. Article 50 covers defined situations and includes role/context questions and exceptions. The screen documents the facts and recommends action only where supported; ambiguity is held for professional review.
Do you need access to our tools?
No. The pilot uses a data-minimized questionnaire, public links and redacted screenshots. DiscloseReady does not need credentials or direct system access.
What pauses delivery?
Missing intake, sensitive data that should not have been shared, unclear authority, high-risk or biometric use cases, and requests for legal certification pause the work.
Start with your actual AI use cases.
Use the secure form to share the team size, the roles using AI, and a short description of the workflows you want documented. Do not include passwords, private prompts, customer datasets, or sensitive personal data.
First responseWe confirm fit, exclusions, inputs and the written scope.
InvoiceThe five-day delivery clock starts after complete intake and cleared payment.
ContactOpen the secure inquiry form