Ensuring compliance with the AI Act: Provisions on AI literacy and prohibited systems apply as of 2 February 2025

At a glance

• Prohibitions on certain AI systems and AI literacy requirements take effect from 2 February 2025.
• Article 4 of the AI Act (Act) mandates that individuals involved in artificial intelligence (AI) systems must possess sufficient AI literacy, including skills and knowledge to use AI effectively and understand its risks and opportunities. 
• Article 5 bans manipulative techniques, social scoring, and emotion recognition, except for specific medical or safety purposes. Violations can result in severe administrative fines.

The Act came into force on 2 August 2024. While the main part of the Act, including the provisions on high-risk systems, applies from 2 August 2026, prohibitions on certain AI systems and requirements on AI literacy will take effect as early as 2 February 2025.

AI literacy

Article 4 of the Act requires providers and operators of AI systems to ensure that individuals involved in the operation and use of these systems possess sufficient AI literacy. This includes the skills, knowledge, and understanding necessary to use AI systems effectively and to understand their opportunities, risks, and potential harms.

The regulation aims to ensure an adequate level of health and safety protection for individuals affected by AI systems. The content of training must be tailored to the recipients, their roles, and the specific AI application. For example, developers of high-risk AI systems need more extensive training than users of low-risk, standardised AI systems.

To ensure AI literacy, employers can introduce the following strategies:

• Company-wide guidelines for AI use.
• Ongoing training to maintain AI competency.
• Voluntary Codes of Conduct to promote AI competency.

Violating Article 4 of the Act does not subject an employer to a fine or penalty. However, if damage occurs as a result of incorrect use of an AI system or an inadequate risk assessment, this could be interpreted as a violation of the employer's general duty of care.

Prohibited AI practices

Article 5 of the Act prohibits certain AI practices, such as manipulative techniques, social scoring, and emotion recognition, unless used for specific medical or safety purposes. Failure to comply with the prohibition on AI practices set out in Article 5 can result in severe administrative fines.

You can access the Act here.

For details of the regulation’s applicability and scope, please see our GENIE news item.