What is AI literacy under Article 4 of the EU AI Act

AI literacy under Article 4 of the EU AI Act refers to a company’s employees and all individuals using AI on its behalf developing knowledge and skills to understand artificial intelligence systems, use them responsibly, and critically evaluate their outputs and risks. This is not a one-off seminar or a formal certificate - it involves knowledge and skills whose development a company must support, taking into account the AI tools it uses, the context in which they are used, and the people they affect. The EU AI Act (Regulation (EU) 2024/1689) establishes this obligation in Article 4, which has applied since 2 February 2025 and covers virtually all companies using AI at work, including the smallest Lithuanian micro-enterprises. Measures must be tailored to the knowledge, experience, training, context, and affected persons, but the Act does not require guaranteeing a specific or sufficient individual level.

For a small business, AI literacy starts with three simple things: clear rules, practical training, and evidence that employees understand the risks. A general seminar or certificate without substantive content is not enough.

What does “AI literacy” mean under the legal act?

The official definition is provided in Article 3(56) of the EU AI Act. AI literacy refers to the skills, knowledge, and understanding that enable providers, deployers, and affected persons to consciously deploy AI systems and understand the opportunities they offer, the risks they pose, and the potential harm they may cause.

In practice, this covers four aspects:

• Understanding what AI is and how it works - so that an employee knows that a tool like ChatGPT generates probable text rather than verifying facts.
• Ability to identify risks - such as incorrect outputs (so-called “hallucinations”), bias, personal data leaks, or disclosure of confidential information.
• Ability to use the tool as intended - knowing when it can be trusted, when human verification is mandatory, and what should never be entered into a conversation.
• Critical evaluation of outputs - ensuring that a human, not a machine, makes the final decision.

Importantly, Article 4 does not prescribe a single mandatory course or exam format, nor does it require specific certificates, a fixed set of documents, a set number of training hours, an external trainer, an audit, or a dedicated AI officer. It requires measures to support the development of AI literacy appropriate to the role and context. A bookkeeper may need less technical knowledge, while someone configuring an AI tool for customer service may need more. Our prepared documents and knowledge-check quiz are recommended practical records, but they are not mandatory legal minimums or a guarantee of compliance.

Who does Article 4 apply to and when?

The obligation under Article 4 applies to AI system providers (those who create AI) and deployers (those who use AI in their operations). Most Lithuanian small businesses are deployers - they do not create their own models but use tools like ChatGPT, Microsoft Copilot, image generators, AI accounting tools, or chatbots. For more details on who the regulation applies to, see who the EU AI Act applies to.

Key dates to note:

Thus, the AI literacy obligation is already in effect - it is the first part of the EU AI Act to take effect. There is no need to wait until 2026. If your team uses any AI tool at work, the obligation already applies to you. Supervision and enforcement of the regulation begin on 2 August 2026.

How is AI literacy different from basic training?

Many companies mistakenly believe that simply showing employees “how to write prompts for ChatGPT” once is sufficient. This is not AI literacy in the legal sense. The difference is fundamental:

• Basic training teaches how to use the tool more effectively - better prompts, faster work.
• AI literacy teaches how to use the tool responsibly and safely - understanding risks, knowing limits, recognising when AI makes mistakes, and being aware of company rules.

Additionally, Article 4 requires that literacy be tailored to the context - the technical knowledge of employees, their roles, and the people affected by AI outputs. For more on specific requirements, see EU AI Act Article 4 requirements, and for what a micro-enterprise realistically needs, see AI literacy requirements for micro-enterprises.

What exactly should a small business do?

Good news: for a micro-enterprise (1 - 30 employees), compliance is not complicated and does not require a lawyer. The practical minimum looks like this:

1. Inventory your AI tools. List which AI tools your team uses - ChatGPT, Copilot, Gemini, Canva AI features, AI accounting tools, etc. This is your AI tool register.
2. Prepare an AI usage policy. Clear rules on what can and cannot be entered into AI, when human verification is required, and how to handle personal data and confidential information. See AI usage policy in a company.
3. Train your employees. Organise AI literacy training tailored to your team’s roles. More details at AI literacy training for employees.
4. Test knowledge and document it. A knowledge check with certificates serves as evidence that you have taken measures. How to do this is explained in how to measure AI literacy.
5. Store the evidence. Dates, participant lists, training content, and certificates - all of this forms your compliance record.

nebedirbam.lt automatically prepares all these documents - AI usage policy, AI tool register, compliance record, employee AI guide, and AI literacy training with knowledge checks and certificates - for a fixed €149 + VAT (€299) price, without the need for a lawyer or consultant. For those who want to start from the basics, there are also free AI literacy training. These documents are recommended practical records but are not mandatory legal minimums or a guarantee of compliance. The free path is for one signed-in participant; team training is included in the paid documentation package.

Why is AI literacy important even without the threat of fines?

Many ask: what happens if I do nothing? Fines for violations of the EU AI Act can reach millions of euros or a percentage of annual turnover - see EU AI Act fines and risks. However, the practical benefits for a small business are often more important than the fines themselves:

• Protects against real harm. An employee who understands the risks will not upload a client’s personal data into a public chatbot and will not send a client an incorrect, AI-generated calculation.
• Reduces GDPR violation risks. AI and personal data are closely linked - see AI and personal data (GDPR). It is important to note that GDPR does not impose a blanket ban on using AI - it requires assessing lawful basis, data minimisation, using approved tools, and implementing safeguards. There is no general prohibition on using AI with personal data, but a proper assessment must be conducted.
• Stops “shadow AI”. When there are no rules, employees use AI secretly and uncontrollably - see shadow AI in a company.
• Builds trust. Clients and partners are increasingly asking how companies manage AI.

Common misconceptions about AI literacy

For clarity, here are a few myths that mislead small businesses:

• “We only use ChatGPT, so it doesn’t apply to us.” False. Any use of AI at work falls under the scope of Article 4. For guidance on safe use, see ChatGPT usage in a company.
• “We need to wait until 2026.” False. The AI literacy obligation has applied since 2 February 2025. Supervision and enforcement begin on 2 August 2026.
• “Signing a document is enough.” False. Real training and knowledge checks are required, not just a piece of paper.
• “It’s expensive and requires a lawyer.” False. For micro-enterprises, the process is standardised and can be fully automated.

More misconceptions are discussed in common AI literacy mistakes.

Brief summary

AI literacy under Article 4 of the EU AI Act is your team’s ability to understand, use responsibly, and critically evaluate artificial intelligence. The obligation has applied since 2 February 2025 to all companies using AI at work, including micro-enterprises. Useful practical measures include: clear rules (AI usage policy and tool register), employee training with knowledge checks, and documented evidence. These documents are recommended but not mandatory legal minimums.

For more information on the EU AI Act and AI literacy, see the list of all articles.

Sources: Official text of the EU AI Act and European Commission guidelines - ai-act-service-desk.ec.europa.eu/en/ai-act/article-4 and digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai.