AI Literacy

Article 4 introduces a universal AI literacy obligation that applies to both providers and deployers, regardless of the risk tier of their AI systems. This obligation became applicable from 2 February 2025 (6 months after entry into force) and is one of the few obligations that applies equally across all AI tiers.

Scope and Content of the Literacy Obligation

Article 4 requires providers and deployers to take measures to ensure a sufficient level of AI literacy of their staff and other persons dealing with AI systems on their behalf. The article specifies that literacy measures must be tailored to the technical knowledge, experience, education, and training of the personnel concerned, and must take into account the context in which the AI system will be used. This means that a data scientist operating a high-risk AI system needs deeper literacy than a customer service agent using an AI-assisted tool.

Practical Implementation

The regulation does not prescribe a specific training curriculum or certification scheme. Organisations have flexibility in how they implement AI literacy, but must be able to demonstrate that literacy measures are proportionate to the role and the AI system involved. Common implementation approaches include role-based training programmes, AI onboarding modules for new staff, regular awareness updates as systems change, and documented literacy assessments for staff operating high-risk AI systems.

• Article 4 applies from 2 February 2025 -- earlier than most other high-risk obligations
• Both providers AND deployers carry this obligation independently
• No specific certification is required -- internal training programmes are sufficient
• Literacy requirements should scale with the risk level and complexity of the AI system
• Documentation of literacy measures supports Article 17 quality management system requirements