Samenvatting

The use and adoption of Artificial Intelligence (AI) in our society is growing rapidly (Maslej 2025). Alongside this growth, the importance of specific ‘competencies’ that users need to effectively interact with and critically evaluate AI is becoming increasingly important (Long en Magerko 2020). This set of skills and awareness is referred to as AI literacy. It is meant to ensure the safe and effective deployment of AI, particularly given the risks tied to using AI systems (ex. ‘hallucinations’ and bias). The principle of AI literacy is reflected in Article 4 of the EU AI Act, which states that both providers and deployers of AI systems must ensure that their personnel and other persons dealing with the operation and use of AI systems possess a sufficient level of AI literacy. This obligation requires a continuous, iterative process involving both technical skills and risk awareness, rather than a static achievement (Recital 20 AI Act; Engelfriet 2025). Recognizing the complexities of this mandate, the Digital Omnibus Proposal of November 2025 is an important development (European Commission 2025b). Its purpose is to simplify several EU regulations, including parts of the AI Act. A key element of the proposal clarifies the role of the European Commission and Member States in ensuring AI literacy. This ‘simplification effort’ acknowledges the ambiguities in the AI Act, especially concerning AI literacy. In practice, AI literacy gains particular weight given the global gap between AI adoption and effective use (Micheletti et al. 2025, 14). Although over 75% of organisations worldwide utilise AI (McKinsey 2025), 74% struggle to scale value primarily due challenges related to people and processes rather than the technology itself (Bellefonds et al. 2025). This emphasizes the importance of AI literacy to realise effective and responsible use of AI. Within Member States, there are variations in the ‘maturity’ and adoption of AI literacy measures. Some implement nation-wide integration programs (e.g., Estonia’s AI Leap), while others primarily rely on informative reports. This variation reflects wider challenges in how the current obligation of AI literacy is governed and enforced. It is also connected to the difficulty of interpreting the AI Act in practice (European Commission 2025a, 74). Even though the European Commission provides information and guidance, it remains unclear what level of AI Literacy is sufficient to be compliant and how this could be demonstrated (Engelfriet & Takens 2025, 92-94). This paper examines the notion of AI literacy in the light of Article 4 of the AI Act. By offering insights into the definition and legal status of the AI literacy requirement, the paper aims to contribute to the current discussion concerning the legal certainty this obligation offers. Furthermore, the paper highlights the differences between ‘maturity’ and adoption of AI literacy across selected Member States and identifies the governance and enforcement challenges that influence this variation. Finally, the paper offers insights into how AI literacy can be embedded practically to ensure compliance with the objectives of the AI Act while also supporting the safe and effective use of AI.