Article 4 of the EU AI Act, which took effect on February 2, 2025, mandates that all EU businesses ensure their workforce possesses sufficient AI literacy to work with or be affected by AI systems. The requirement is binding, not optional. AI literacy encompasses understanding how AI works, evaluating information input in accordance with data protection law, using AI in compliance with ethical standards, and utilizing technology with necessary risk awareness. For high-risk AI systems (those affecting healthcare, law enforcement, critical infrastructure, and education), comprehensive training is mandatory for all employees involved in development, implementation, or monitoring. Full compliance is due by August 2, 2026 (or August 2027 for Article 6, Section 1, pending omnibus procedure outcomes).
Who it affects: All employees in EU member states working with or subject to AI systems. Employers are obligated to provide training. The high-risk designation extends mandatory comprehensive training to sectors where AI decisions directly impact people’s rights and safety, reaching employees in critical public and private sectors.
What is notably missing: The regulation does not specify minimum training hours, curriculum content standards, or assessment methods to verify competency. No funding mechanism is established to support SMEs in providing required training. No enforcement penalties are explicitly detailed for non-compliance. The regulation does not mandate training for gig workers, self-employed persons, or informal labour. No provision for AI ethics training going beyond “using AI responsibly” is mandated in the standard framework, though high-risk systems do require understanding of ethical and legal implications. No requirement to update training as AI systems and risks evolve.