Legal status June 2026UAIV consultation closed (1 June) · Digital Omnibus: political agreement 7 May 2026Legal status ›
NL · EN
For organisations

For organisations.

Article 4 of the EU AI Act obliges every organisation that uses or provides AI to ensure a demonstrable level of AI literacy among its staff. This overview shows what the law requires, how SAIG certification aligns with the Dutch Data Protection Authority's four-step plan, and how to arrive at a workable approach.

What Article 4 requires

The legal text, in one paragraph.

Providers and deployers of AI systems take measures to ensure, to their best extent, a sufficient level of AI literacy among their staff and other persons dealing with the operation and use of AI systems on their behalf, taking into account their technical knowledge, experience, education and training and the context the AI systems are to be used in.

State of play — June 2026Article 4 has applied since 2 February 2025; enforcement of the operator obligations starts on 2 August 2026. The obligations for stand-alone high-risk AI (Annex III) were postponed to 2 December 2027 under the Digital Omnibus agreement. The Dutch AI Act Implementation Act (UAIV) was open for consultation until 1 June 2026; the legislative process is ongoing. On 7 May 2026 the Council and Parliament reached a provisional political agreement on the Digital Omnibus; Article 4 was not marked as definitively amended therein and remains in force.

The Dutch DPA four-step plan

Identify, set goals, implement, evaluate.

STEP 01

Identify

Map every AI system in use, including shadow AI. Document who works with it, in what context and at what risk.

STEP 02

Set goals

Define measurable, role-specific learning objectives, prioritising the highest risks. Secure executive ownership and budget.

STEP 03

Implement

Roll out training, communicate policy and embed AI policy in onboarding and recurring mandatory modules.

STEP 04

Evaluate

Measure knowledge, behaviour and outcomes through assessments, audits and management reporting. Schedule an annual review.

Procurement and HR policy

From learning outcome to procurement criterion.

SAIG certification is not only a learning outcome — it is a procurement criterion. In their HR policy, onboarding requirements and supplier contracts, organisations can set out which SAIG level applies per function or role. In tenders, a verifiable SAIG certificate provides demonstrable evidence to clients and supervisors that staff meet Article 4. The public register enables two-step verification — without making personal data public. AI literacy is therefore not only trained but also assured and made auditable throughout the procurement chain.

Which SAIG level for which role

Level per role — a starting point.

A role-specific approach calls for differentiation. The table below offers a starting point — the final choice depends on risk class, sector and the substance of the role.

Role groupSAIG levelRationale
Broad group of staff
administrative, support
Awareness Badge or BasisAwareness and safe basic use in low-risk contexts.
Frequent AI users
knowledge workers, consultants
SAIG-PractitionerStructural application in work processes, with attention to verification and risk.
Deployers
responsible users
Practitioner or AdvancedResponsibility for system deployment under the AI Act.
Developers and procurers
providers
SAIG-AdvancedDeeper technical and legal knowledge for lifecycle responsibility.
Executives, AI Officers, compliance, lawyersSAIG-AdvancedGovernance and assurance responsibility at organisational level.
HR and trainersPractitioner or AdvancedGuiding and training others requires an overview of all levels.

Sectoral approach

Ten main sectors, one horizontal foundation.

Article 4 applies to every sector, but risk profile and supervision differ. For each main sector there is a separate page covering what is at play, the relevant legislation and the appropriate level per role.

Healthcare & wellbeing

GDPR art. 9, MDR/IVDR and IGJ supervision; AI in diagnostics and reporting.

Lees meer →

Government

Algorithm register, IAMA/FRIA and high-risk in public services.

Lees meer →

Financial

Credit scoring and insurance high-risk; DNB, AFM, DORA.

Lees meer →

Education

Admission, assessment and proctoring high-risk; emotion recognition prohibited.

Lees meer →

Legal

Confidentiality, professional ethics and the NOvA recommendations.

Lees meer →

Professional services

HR AI high-risk; AFM and NBA tightening supervision.

Lees meer →

ICT, media & creative

Highest adoption; Article 50, the DSA and copyright.

Lees meer →

Industry

Annex I via product safety; Machinery Regulation and NIS2.

Lees meer →

Construction, energy & utilities

Critical infrastructure high-risk; NIS2 and smart grid management.

Lees meer →

Retail & logistics

HR AI high-risk, facial recognition restricted, transparent pricing.

Lees meer →

Next step

Schedule an orientation call.

SAIG offers exploratory orientation calls for organisations seeking to implement Article 4. In the call we discuss your context, the scale of your AI use, and the appropriate SAIG route.

Schedule an orientation call →