Article 4 of the AI Act requires every organisation that uses or develops AI to demonstrate AI literacy among employees — aligned with their role, responsibilities, and the risk class of the AI systems they work with. The deadline of 2 February 2025 has passed. Most Dutch HR directors still do not have a systematic literacy plan. This article gives you the structure.
What AI literacy legally requires
Article 4 is deliberately technology-neutral: "Providers and deployers of AI systems shall take measures to ensure, to the extent reasonably foreseeable, a sufficient level of AI literacy among their staff." The ACM and other national supervisors expect demonstrable effort in 2026 audits: documentation of the literacy plan, evidence of training, and differentiated programmes by role and risk level.
What does not comply: a generic e-learning module on AI that all employees must complete. What does comply: a differentiated programme that accounts for the AI systems specific roles work with, the risk class of those systems, and the employee's specific decision-making responsibilities.
The three-layer literacy model
AlphaIndigo applies a three-layer model for organisation-wide AI literacy:
Layer 1: Organisation-wide baseline literacy. All employees understand what AI is, which AI systems the organisation deploys, what the risks are for their work, and how to report unwanted or problematic AI output. This is the minimum compliance baseline. Duration: 2–4 hours, online. Certification recommended for audit trail.
Layer 2: Role-specific AI competencies. Employees who use AI tools daily in their work (customer contact, credit assessment, HR selection, financial advice) develop specific competencies: critical evaluation of AI outputs, recognition of bias in AI recommendations, escalation paths for uncertain AI decisions, and documentation requirements. Duration: 8–16 hours per cohort. Blended learning recommended.
Layer 3: AI governance leadership. Decision-makers and governance owners understand the regulatory context (AI Act, GDPR Article 22, sector-specific requirements), risk classification of AI systems, their personal accountability for AI decisions, and governance frameworks for responsible AI use. Duration: 1–2 days. Executive programme, strongly recommended individually.
The competency framework
For layers 2 and 3, a competency framework is essential. AlphaIndigo bases its approach on five AI competency domains:
1. AI understanding. Employees know how the AI systems in their work function — not technically, but conceptually. They understand what the input is, how the output is produced, and what the limitations are.
2. Critical use. Employees can assess AI outputs for quality, consistency, and reliability. They know when to override and how to document it.
3. Privacy and security. Employees know which data they may and may not enter into AI systems. They understand the risks of shadow AI and unapproved tools.
4. Ethical judgement. Employees recognise situations where AI use is ethically or legally problematic. They know how to escalate.
5. Continuous learning. AI systems evolve quickly. Employees and the organisation need a mechanism for ongoing literacy renewal.
The HR director as governance owner
In Dutch enterprises, the HR director is typically the functional owner of AI literacy — even when technical implementation sits with IT. This means that as HR director you fulfil three roles: architect of the literacy plan, enforcer of compliance requirements, and internal adviser to the board on the human dimension of AI governance.
The most effective HR directors AlphaIndigo supports begin with an AI systems inventory together with IT and Legal: which AI systems are active, which roles use them, what is the risk class, and which Article 4 obligations follow. Only on the basis of this inventory is the literacy plan built.
The AI Academy: three tracks
The AI Academy from AlphaIndigo is designed as a modular programme aligned with the three-layer model. Track 1 (baseline literacy) is online, scalable, and tailored to your organisation context — including your specific AI systems and sector regulation. Track 2 (role-specific competencies) is delivered blended per function cohort, with sector-specific case studies (DNB expectations for financial services, IGJ for healthcare, ACM for telecom operators). Track 3 (AI governance leadership) is an executive programme for board members, CISOs, CDOs, and HR directors themselves. All tracks deliver certification for the audit trail — and are fully documented in accordance with Article 4 requirements.