AI literacy is the set of skills, knowledge and understanding that enables people to use artificial intelligence systems in an informed way, understand their capabilities and recognize the risks associated with their use.
Since February 2, 2025, Article 4 of the AI Act has established an AI literacy obligation for providers and deployers of AI systems.
But there is an important update.
Article 4 was amended in July 2026 by Regulation (EU) 2026/1744.
The obligation remains in place, but organizations no longer have to ensure that every individual reaches a specific or “sufficient” level of AI literacy.
They must take measures that support the development of these capabilities while taking into account knowledge, experience, education, training and context.
The European Commission explains these obligations in its official questions and answers on AI literacy.
If you are looking for a broader explanation of the regulation, see our complete guide to the EU AI Act.
The Essentials of AI Literacy
- Article 4 has applied since February 2025.
- AI literacy remains a legal obligation for providers and deployers.
- The 2026 reform removed the requirement to guarantee a specific level of knowledge for each individual.
- There is no single mandatory course.
- There is no specific certificate required by Article 4.
- Measures should be adapted to the person’s role, knowledge and the system being used.
- A company whose employees use ChatGPT is expressly included in one of the examples provided by the European Commission.
- Organizations may keep internal records of training and other initiatives as evidence of the measures they have taken.
If you first want to identify which areas of the AI Act your organization should review, use the Founderz AI Act diagnostic.
What Is AI Literacy?
The regulation defines AI literacy as the skills, knowledge and understanding that allow people to make an informed deployment of artificial intelligence systems and gain awareness of:
- opportunities;
- risks;
- and potential harm.
AI literacy is not simply about learning how to write prompts.
A person may use ChatGPT every day and still need training if they do not understand:
- when the model may be wrong;
- how to interpret its output;
- what information they should not enter;
- what biases may appear;
- when an answer needs to be verified;
- when a decision requires human oversight.
Understanding how AI works at the level appropriate to a person’s role is therefore only one part of AI literacy.
The professional, ethical, legal and organizational implications of using the technology also matter.
What Does Article 4 of the AI Act Currently Require?
Article 4 of the AI Act requires providers and deployers to take measures to support the development of AI literacy among:
- their staff;
- and other people involved in operating and using AI systems on their behalf.
The measures should take into account:
- technical knowledge;
- experience;
- education;
- training;
- context of use.
The Commission and Member States must also support these efforts, particularly for small and medium-sized businesses.
What Changed in July 2026?
The original wording used the concept of a “sufficient level” of AI literacy.
The Digital Omnibus amended Article 4.
The obligation still exists, but the new framework clarifies that providers and deployers do not have to guarantee a specific level for each individual.
There is also no general obligation to formally test every employee to establish a defined level of knowledge.
The Commission expressly confirms this in its updated questions and answers on Article 4.
Who Does the AI Literacy Obligation Apply To?
Article 4 is not limited to technology companies.
It primarily covers:
Providers
Organizations that develop certain AI systems and market or place them into service under their own name.
Deployers
Organizations that use artificial intelligence systems under their authority as part of a professional activity.
Other People Who Use AI on Behalf of the Organization
Depending on the context, this may include:
- employees;
- contractors;
- service providers;
- other professionals.
This does not mean everyone needs exactly the same training.
Does a Company That Uses ChatGPT Need AI Literacy?
It may.
The European Commission expressly uses the example of an organization whose employees use ChatGPT to write advertising copy or produce translations.
Its response is that those people should understand specific risks such as hallucinations.
Other everyday uses may include:
- Microsoft Copilot for summarizing meetings;
- generative AI for drafting documents;
- systems for analyzing information;
- customer service assistants;
- image-generation tools;
- internal automations.
To examine these cases from the employee’s perspective, see AI Act for Employees: What You Need to Know if You Use AI at Work.
Is an AI Literacy Course Mandatory?
There is no single mandatory course established by the AI Act.
The Commission states that there is no one-size-fits-all solution.
Depending on the context, an AI literacy program may combine:
- AI training;
- online courses;
- in-person sessions;
- internal guidance;
- policies;
- tool documentation;
- workshops;
- department-specific training;
- practical exercises;
- regular updates.
In some cases, simply providing a tool’s user manual may not be enough.
The aim is for the person to have capabilities appropriate to the system they use.
Is a Certificate Required?
No.
The European Commission expressly confirms that Article 4 does not require a specific certificate.
A training certificate may demonstrate that a person completed a particular program, but it does not by itself demonstrate comprehensive compliance with the AI Act.
The company needs to assess the overall measures it has taken and whether they are appropriate to the context.
What Should an AI Literacy Program Include?
1. Understand Which AI the Organization Uses
The company should know:
- which systems it has;
- who uses them;
- what tasks they are used for;
- what opportunities they present;
- what risks they create.
2. Determine the Company’s Role
Acting as a provider is not the same as acting as a deployer.
3. Assess the Risk
The depth required depends on the system and the use case.
Using AI to produce a first draft of some text is not equivalent to using it for recruitment or for a decision that directly affects people’s rights.
4. Adapt the Training
The organization should consider:
- experience;
- knowledge;
- responsibility;
- previous training;
- context;
- the specific technology.
5. Include Legal and Ethical Issues
A robust program may cover:
- fundamental rights;
- data protection;
- confidentiality;
- intellectual property;
- bias;
- hallucinations;
- security;
- human oversight;
- responsible use.
Example AI Literacy Levels
| Level | Audience | Indicative Content |
|---|---|---|
| Foundation | People starting to use AI | Basic concepts, opportunities, limitations, safety and verification. |
| Applied | Employees who use AI regularly | Prompting, output evaluation, data, bias and good practices. |
| Specialized | Managers, responsible teams or technical teams | Governance, risk, policies, oversight and applicable obligations. |
| High risk | People managing high-risk AI systems | Specific training, human oversight, risks and procedures. |
This table is not an official classification under the regulation. It is a practical way to organize AI literacy programs.
What Documentation Should You Keep?
The Commission states that an organization may maintain an internal record of training and other guidance initiatives.
For example:
- what activity took place;
- who participated;
- date;
- content;
- version of the material;
- systems covered;
- subsequent reviews.
A reasonable organizational structure may include:
- an inventory of AI systems;
- a map of users and roles;
- a needs assessment;
- training measures;
- materials;
- participation records;
- updates.
There is no single mandatory template.
European Commission Repository of AI Literacy Practices
The Commission maintains a Living Repository of AI literacy practices.
It contains real initiatives used by companies and public administrations through:
- e-learning;
- in-person training;
- boot camps;
- internal programs;
- collaboration with educational organizations.
It is a useful source for identifying good practices.
However, the Commission itself warns that replicating a practice from the repository does not automatically create a presumption of compliance with Article 4.
That is why measures need to be adapted to the organization’s actual use of AI.
Benefits of AI Literacy for Companies
More Informed Use of AI
A trained employee is better able to identify when an apparently correct answer needs verification.
Better Adoption
The organization can move from improvised use of tools to defined use cases.
Less Unnecessary Exposure of Information
Training helps people understand which data they can enter and which information requires additional controls.
Stronger Human Oversight
AI literacy helps ensure that human review is meaningful rather than merely formal.
Better Collaboration
Technical teams, HR, legal, marketing, operations and management can use a common language.
More Responsible Innovation
Understanding opportunities and risks makes it easier to integrate AI without allowing the speed of adoption to become a source of problems.
AI Literacy and High-Risk Systems
The general obligation under Article 4 does not replace other obligations in the regulation.
Certain high-risk systems have specific requirements relating to the training of people responsible for human oversight.
Therefore:
general AI literacy under Article 4
and
specific training for operating certain high-risk systems
are related but separate concepts.
When Is Article 4 Compliance Supervised?
The obligation began to apply on February 2, 2025.
Since August 2026, supervision and enforcement rules have been active.
Supervision of this general obligation does not fall directly to the AI Office. It primarily lies with national market surveillance authorities.
Enforcement must follow a proportionate approach and consider circumstances such as:
- nature of the infringement;
- severity;
- negligence;
- measures taken;
- consequences.
It is therefore incorrect to claim that failing to provide a particular course automatically results in the maximum penalty.
How to Start an AI Literacy Program
A practical route:
- Inventory the tools being used.
- Identify who uses them.
- Analyze the use cases.
- Assess risks and required knowledge.
- Create training measures for different profiles.
- Document those measures.
- Update the program regularly.
If you first want to identify which areas of the regulation may affect your organization, use the Founderz AI Act diagnostic.
Frequently Asked Questions About AI Literacy
What Is AI Literacy?
It is the set of skills, knowledge and understanding that allows people to use artificial intelligence systems in an informed way and understand their opportunities, risks and potential harm.
Is AI Literacy Mandatory?
Article 4 requires providers and deployers to take measures supporting its development among people who use systems on their behalf.
Is It Still Mandatory to Guarantee a “Sufficient Level”?
No. The 2026 amendment removed the obligation to guarantee a specific level of AI literacy for each individual.
Is a Course Mandatory?
There is no single course imposed by the EU.
Do Employees Have to Be Certified?
There is no mandatory certification under Article 4. Companies may maintain internal records of training and other initiatives.
Does It Affect a Company That Only Uses ChatGPT?
It can. The Commission itself uses this as an example of a company that needs to consider specific risks associated with the use of AI.
Does the Training Have to Be the Same for Everyone?
No. It should take into account knowledge, experience, education, role and context.
Can It Include Contractors?
Yes. Depending on the circumstances, other people using AI systems on behalf of the organization may fall within the obligation.
Does Training Guarantee Compliance With the AI Act?
No. Training may be an important measure, but no single training program guarantees compliance with all applicable obligations.
Where Can I Understand the Full Regulation?
See our complete guide to the EU AI Act.
AI Literacy Training for Companies
The goal is not to turn every employee into a technical specialist.
It is to develop capabilities appropriate to how each person uses artificial intelligence.
The Founderz AI Literacy program provides practical training for professionals and teams that need to develop these skills.
It can form part of an AI literacy strategy alongside policies, processes, guidance and internal controls.
If you are still unsure what your organization needs, start with the Founderz AI Act diagnostic.
Disclaimer: No specific training program by itself constitutes a guarantee of compliance with Article 4 or the AI Act. This article is provided for informational and educational purposes and does not replace specialist legal advice.