AI Literacy under Article 4 AI Act: What Providers and Deployers Must Do after the 2026 Amendment

AI literacy under Article 4 of the AI Act is the duty of providers and deployers to take measures that support the AI literacy of their staff and of other persons dealing with AI systems on their behalf. Since 27 July 2026, Article 4 has three paragraphs: the duty itself, support from the Commission and the Member States, and recommendations of the European Artificial Intelligence Board. The duty does not require any provider or deployer to guarantee a specific level of AI literacy of any individual.

Article 4 reaches providers and deployers of every AI system, not only of high-risk systems. The EU AI Act is Regulation (EU) 2024/1689, in force since 1 August 2024 and amended by Regulation (EU) 2026/1744, the Digital Omnibus on AI, which replaced Article 4 with effect from 27 July 2026. Article 4 itself has applied since 2 February 2025, so the amendment rewrote a duty that organisations already had to meet.

Compiled by the AIActGuide.net Editorial Team from the Official Journal text; how we check every statement: methodology.

What Is AI Literacy? The Legal Definition and Its Reach

AI literacy in the AI Act means the skills, knowledge and understanding that let providers, deployers and affected persons make an informed deployment of AI systems (Article 3(56)). The definition also covers awareness of the opportunities and risks of AI and of the harm it can cause, measured against the rights and obligations each person has under the act. The term carries this legal sense only: it names what Article 4 measures aim at, not a school subject or a general education curriculum.

The Article 3(56) Definition and Whom It Names

Article 3(56) defines AI literacy for providers, deployers and affected persons, while Article 4 places the duty only on providers and deployers for their staff and people acting on their behalf. The definition reads: "'AI literacy' means skills, knowledge and understanding that allow providers, deployers and affected persons, taking into account their respective rights and obligations in the context of this Regulation, to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause". Regulation (EU) 2026/1744 left Article 3(56) unchanged.

The definition is wider than the duty. Recital 20 of Regulation (EU) 2024/1689 says AI literacy should equip providers, deployers and affected persons with the notions needed to make informed decisions about AI systems, yet no Article 4 measure is owed to affected persons. Among the AI Act definitions in Article 3, point (3) defines the provider and point (4) the deployer, the two roles that carry the duty.

The Article 4 Text in Force since 27 July 2026

Article 4 has three paragraphs since 27 July 2026, when Article 1 point (5) of Regulation (EU) 2026/1744 replaced the one-paragraph text of 2024. The EU AI Act full text carries the new wording in the consolidated version as at 27 July 2026. The AI Act Service Desk marks all three paragraphs as amended on its Article 4 page, checked on 24 October 2026.

Paragraph 1: The Duty of Providers and Deployers

Article 4(1) obliges providers and deployers of AI systems to take measures to support the development of AI literacy, calibrated to the people concerned and the context of use.

"Providers and deployers of AI systems shall take measures to support the development of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf, taking into account their technical knowledge, experience, education and training and the context the AI systems are to be used in, and considering the persons or groups of persons on whom the AI systems are to be used. This obligation does not require providers or deployers to guarantee any specific level of AI literacy of any individual." (Article 4(1), as replaced by Regulation (EU) 2026/1744)

The paragraph names three calibration factors: the technical knowledge, experience, education and training of the people concerned, the context of use, and the persons or groups on whom the systems are used. Its second sentence removes any guarantee, so no provider or deployer answers for the level a single person reaches.

In our reading, the paragraph sets a duty of conduct rather than a duty of result: the test is whether the measures fit the staff, the AI systems and their context, not what any one employee ends up knowing.

Paragraph 2: Support from the Commission and the Member States

Article 4(2) requires the Commission and the Member States to support providers and deployers, in particular SMEs, and the Commission to publish practical examples on the single information platform.

"The Commission and the Member States shall support and facilitate the efforts of providers and deployers of AI systems, in particular SMEs, in fulfilling their obligation under paragraph 1 of this Article. For that purpose, the Commission shall publish practical examples of how to comply with that obligation on the single information platform referred to in Article 62(3), point (b)." (Article 4(2))

That platform is the AI Act Service Desk. No practical examples under Article 4(2) are published on the AI Act Service Desk as of 24 October 2026, and the Commission AI Literacy Q&A says the Commission "will publish" them. The support adds to the Article 62 measures that the guide to the AI Act for SMEs sets out; Article 62 applies since 2 August 2026.

Paragraph 3: Recommendations of the AI Board

Article 4(3) tasks the European Artificial Intelligence Board with adopting recommendations that support the Commission and the Member States in promoting AI literacy.

"The Board shall adopt recommendations, taking into account European competence frameworks, to support the Commission and Member States in the promotion of AI literacy required under paragraph 1, including by setting out common objectives." (Article 4(3))

The recommendations build on an existing task: Article 66, point (f), asks the Board to support the Commission in promoting AI literacy, public awareness and understanding of the benefits, risks, safeguards, rights and obligations in relation to the use of AI systems, and Article 66 applies since 2 August 2025. No Board recommendations under Article 4(3) are adopted as of 24 October 2026; the Commission's AI talent, skills and literacy page, last updated on 13 October 2026, describes them as work in progress. Until they exist, no common objectives at EU level shape the Article 4 measures of any organisation.

How the 2026 Amendment Changed Article 4: Three Versions Compared

The 2026 amendment kept the duty on providers and deployers but replaced the obligation to ensure a sufficient level with one to take measures supporting its development; the proposal to move the duty to the Commission and the Member States was not adopted. The three versions differ in who carries the duty and in the verb that defines it:

Article 4 in Regulation (EU) 2024/1689, COM(2025) 836 and Regulation (EU) 2026/1744.

The new EU rule on AI literacy came through an adopted regulation, not through the proposal. The Digital Omnibus on AI is Regulation (EU) 2026/1744 of 8 July 2026, published in the Official Journal on 24 July 2026 and in force since 27 July 2026. Its Article 1 point (5) states "Article 4 is replaced by the following" and inserts the three paragraphs now in force.

Recital 8 of Regulation (EU) 2026/1744 gives the reason: stringent obligations to ensure a sufficient level of AI literacy would not suit all types of providers and deployers and would add compliance burden, particularly for smaller enterprises. The same recital calls AI literacy "a strategic priority regardless of regulatory obligations and potential sanctions"; it is a recital of the amending regulation, not one inserted into the AI Act. The corrigendum 32026R1744R(01) of 29 September 2026 (OJ L, 2026/90810) concerns only the German version, corrects Article 1 point (8) on Article 6(1b) and does not touch Article 4.

The AI Literacy Section of This Guide: Three Pages and What Each Answers

Article 4 leaves three practical questions to each organisation: what an AI literacy programme covers, whether an AI officer is needed, and what an internal AI usage policy contains.

Article 95(2), point (c), allows voluntary codes of conduct to cover "promoting AI literacy, in particular that of persons dealing with the development, operation and use of AI", so a written AI policy can also form part of such a code.

Who Must Take AI Literacy Measures: Providers, Deployers and Persons on Their Behalf

Article 4 binds providers and deployers of AI systems, so an organisation that places an AI system on the market, puts it into service or uses it professionally carries the duty. The provider vs deployer roles follow Article 3: a provider develops an AI system or has it developed and places it on the market or puts it into service under its own name or trademark, while a deployer uses an AI system under its authority. Importers, distributors and authorised representatives are not named in Article 4(1), although one organisation can hold more than one role.

Staff, Contractors, Service Providers and Clients

The duty covers the provider's or deployer's staff and other persons dealing with the operation and use of AI systems on its behalf. Other persons are, according to the Commission AI Literacy Q&A, people who are not employees but fall broadly under the organisational remit, for example a contractor, a service provider or a client.

The Q&A treats everyday generative AI use as covered. Companies whose employees use ChatGPT to write advertising text or to translate content must comply with Article 4, and their staff should be informed of risks such as hallucination.

Personal Non-Professional Use and Affected Persons

A natural person using an AI system in a purely personal non-professional activity is not a deployer and carries no Article 4 duty (Article 3(4), Article 2(10)). Article 3(4) excludes from the deployer definition any use "in the course of a personal non-professional activity", and under Article 2(10) the act does not apply to the obligations of deployers who are natural persons using AI systems in a purely personal non-professional activity. The provider of the same AI system stays bound by Article 4. Affected persons appear in the Article 3(56) definition but are owed no Article 4 measure; their protection comes from other provisions, such as the Article 50 information duties.

What Providers and Deployers Must Do: The Measures Article 4 Expects

Providers and deployers meet Article 4 with AI literacy measures that fit their role, the risk of their AI systems and the knowledge of the people concerned. Article 4 prescribes no method, and this section, drawn from Article 4 and the Commission AI Literacy Q&A, gives an orientation, not a legal assessment of any individual case.

The Four Considerations in the Commission Q&A

The Commission AI Literacy Q&A names four minimum considerations in its question 2.1, each building on the one before:

• A general understanding of AI within the organisation.
• The role of the organisation, as provider or as deployer of the AI systems concerned.
• The risk of the AI systems that the organisation provides or deploys.
• AI literacy actions built on that analysis, considering the staff's knowledge, experience, education and training, the context of use and the affected persons.

The Q&A is non-binding Commission guidance. According to the Q&A, the Commission plans no dedicated guidelines on Article 4, and AI literacy is to be touched upon in the guidelines on Articles 8 to 15 and Article 25. As its own practice, the Commission describes an internal model built on a Policy to Build and Maintain an AI-skilled Workforce, a Competency Framework for Basic AI Literacy, learning packages for generalists, managers and developers, and an AI Champions network.

What Article 4 Does Not Require

Article 4 sets no certificate, no mandatory training format, no knowledge test and no dedicated officer. The Commission AI Literacy Q&A states five exclusions:

• No certificate: "There is no need for a certificate. Organisations can keep an internal record of trainings and/or other guiding initiatives."
• No mandatory training format: "no strict requirements or mandatory trainings are imposed", although relying only on the instructions for use of an AI system might be ineffective.
• No knowledge test: there is no obligation to measure the AI knowledge of employees.
• No dedicated officer: "No, no specific governance structure is mandated to comply with article 4 of the AI Act." The question of an AI officer turns on this answer.
• No sector rules: the AI Office imposes no sector-specific requirements, although the sector and the purpose of use matter as context.

In our reading, an internal record of the measures taken is what an organisation can show when a market surveillance authority asks how it meets Article 4. What a programme contains and how that record is kept are questions of EU AI Act training under Article 4(1), not of formal certification.

Article 4 distributes duties among five addressees: providers, deployers, the Commission, the Member States and the AI Board. Deployers of high-risk AI systems carry a separate competence duty under Article 26(2), shown in the last row for contrast:

Article 4 and the Competence Duty for Deployers of High-Risk Systems (Article 26(2))

Deployers of high-risk AI systems carry a second, stricter duty: Article 26(2) requires them to assign human oversight to natural persons with the necessary competence, training and authority. The provision for high-risk AI systems reads: "Deployers shall assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support." Article 26(2) applies from 2 December 2027 for Annex III systems and from 2 August 2028 for Annex I systems, the dates on which Chapter III Sections 1 to 3 apply to them (Article 113, third paragraph, point (c), as replaced by Regulation (EU) 2026/1744).

Recital 91 of Regulation (EU) 2024/1689 links the two duties: deployers should ensure that the persons implementing the instructions for use and human oversight have "the necessary competence, in particular an adequate level of AI literacy, training and authority". The Commission AI Literacy Q&A confirms that this training duty for deployers of high-risk systems remains in place after the 2026 amendment. Article 26(2) stays separate: it targets named oversight persons, while Article 4 reaches all staff and other persons dealing with AI systems.

Article 99(4), point (e), lists the obligations of deployers under Article 26 among the infringements for which national penalty rules provide administrative fines of up to EUR 15 000 000 or 3 % of total worldwide annual turnover, whichever is higher. Deployer obligations under Article 26 and human oversight under Article 14 together define what a high-risk deployer owes beyond Article 4.

When Article 4 Applies: From 2 February 2025 to the Amended Text

Article 4 applies since 2 February 2025 under Article 113, third paragraph, point (a), and in its amended wording since 27 July 2026. Point (a) covers Chapter I, Articles 1 to 4: subject matter, scope, definitions and AI literacy. The EU AI Act timeline places these dates among all other application dates of the act.

Before 27 July 2026, Article 4 required providers and deployers to "take measures to ensure, to their best extent, a sufficient level of AI literacy". That wording is history.

The amended Article 4 sets no deadline for the practical examples of paragraph 2 or the Board recommendations of paragraph 3, and neither exists as of 24 October 2026. Both steps depend on the Commission and the AI Board, not on providers or deployers.

Enforcement of Article 4: Supervision, Fines and National Rules

National market surveillance authorities supervise Article 4, from 2 August 2026 according to the Commission's AI Literacy Q&A of 20 October 2026. That supervision sits within the national enforcement framework, and the EU AI Act penalties regime of Chapter XII applies since 2 August 2025. Under Article 70, Member States shall designate these market surveillance authorities.

Private enforcement runs on national law. According to the Q&A, persons harmed by a breach can bring claims under national law, while the AI Act itself creates no criminal offences and no right to compensation.

No EU Fine Amount for Article 4

The AI Act sets no fine amount for Article 4: the provision is not listed in Article 99(4), so any penalty depends on national rules under Article 99(1). Article 99(1) requires Member States to "lay down the rules on penalties and other enforcement measures", and "the penalties provided for shall be effective, proportionate and dissuasive". Deployer obligations under Article 26 sit in the EUR 15 million tier of Article 99(4), point (e), while Article 4 sits in no tier. A national penalty for an Article 4 breach therefore exists only where a Member State has created one.

Germany as a Worked Example: National Supervision without an Article 4 Offence

Germany shows how Article 4 is supervised nationally: the Bundesnetzagentur is the default market surveillance authority under § 2(1) KI-MIG, and § 15 KI-MIG creates no administrative offence for Article 4. The KI-MIG was signed on 22 July 2026 and promulgated in BGBl. 2026 I Nr. 223 of 28 July 2026. Under § 11(1) KI-MIG, the market surveillance authorities hold the powers of Article 14(4) and (5) and Article 16 of Regulation (EU) 2019/1020 as well as the powers under the AI Act.

§ 15 KI-MIG lists the national offences: breaches of certain duties under Articles 21, 27 and 45 and of Article 86(1), with fines of up to EUR 50 000 under § 15(3). Article 4 is not on that list, so a missing AI literacy measure is no administrative offence under the KI-MIG as of 24 October 2026. The guide to the AI Act in Germany sets out the other German authorities and their tasks.

The Bundesnetzagentur's guidance on AI literacy reaches further than the offence list: the authority states that a lack of AI literacy can be regarded as a breach of the duty of care, in particular where damage results, and recommends documenting the measures taken. Under § 5 KI-MIG, the coordination and competence centre KoKIVO at the Bundesnetzagentur offers training and awareness activities and general information.

Official Resources on Article 4: Commission Q&A, Service Desk and Repository

Four Commission sources explain Article 4: the AI Literacy Q&A, the AI Act Service Desk, the AI talent, skills and literacy page and the living repository of AI literacy practices. Each carries its own update date:

Replicating a practice from the living repository "does not automatically grant presumption of compliance with Article 4", in the Commission's words on the repository page. The register of EU AI Act guidelines lists every Commission guideline under the act with its status. Pages dated before 27 July 2026 describe the 2024 wording of Article 4, so their quotations of the duty are history.

Putting Article 4 into Practice: From an AI Inventory to Records

Five steps turn Article 4 into a working routine: an inventory of AI systems, the role per system, the people concerned, the measures and an internal record. The steps follow Article 4(1) and the four considerations of the Commission Q&A, and they give an orientation, not a legal assessment:

1. List every AI system the organisation provides or deploys, including generative AI tools that staff use for text or translation.
2. Assign the role per system: provider, deployer or both.
3. Map the people concerned by role and by their technical knowledge, experience, education and training.
4. Choose measures that fit the role, the risk of each system and the context of use.
5. Record the measures in an internal record and refresh them when AI systems or staff change.

The inventory and the role decide every later step, in Article 4 as in the rest of the act. A programme that combines AI literacy with the risk-class duties is set out in the guide to EU AI Act compliance, written for providers and deployers alike. AI literacy measures cover every AI system, while the other duties attach to specific categories, such as high-risk AI systems or systems under Article 50.

Public support exists for the measures themselves: under Article 62(1), point (b), Member States organise awareness raising and training activities on the application of the AI Act, tailored to the needs of SMEs, start-ups, deployers and, as appropriate, local public authorities. The content of EU AI Act training, from a basic understanding of AI to role-specific modules, remains a choice of each organisation.

Each duty of the act can be traced to its article, its addressee and its date. The EU AI Act checklist sorts the obligations by role and risk class, with Article 4 among the duties that already apply. Updating the inventory whenever a new AI system arrives keeps the internal record current.

AI Literacy under the AI Act: Related Questions

Article 4 connects with Article 2 on territorial scope, Article 113 on application dates, Article 50 on transparency and Article 70 on national authorities.

Do Non-EU Companies Have to Meet Article 4?

Non-EU companies have to meet Article 4 where their AI system is placed on the Union market, used in the Union or its use has an impact on people located in the EU, according to the Commission AI Literacy Q&A.

The territorial reach of the act for organisations outside the Union is explained in the guide to the EU AI Act for non-EU companies. The same measures apply to them as to EU organisations.

Which Other Rules Apply from the Same Date as Article 4?

Chapters I and II of the AI Act apply from the same date as Article 4, 2 February 2025.

The prohibited AI practices of Article 5 apply in two steps: points (a), (b) and (c) to (h) since 2 February 2025, and points (ba) and (bb) together with Article 5(1a) and (1b) from 2 December 2026. Articles 1 to 3 on subject matter, scope and definitions complete Chapter I alongside Article 4.

How Does Article 4 Differ from the Article 50 Transparency Duties?

Article 50 sets transparency duties towards the people who interact with AI systems or are exposed to their output, while Article 4 equips the organisation's own staff and the persons acting on its behalf.

The transparency obligations under Article 50 apply since 2 August 2026 under Article 113, second paragraph, unchanged by Regulation (EU) 2026/1744. Article 50(1) and (2) bind providers, Article 50(3) and (4) bind deployers, and breaches fall in the EUR 15 million tier of Article 99(4), point (g).

Who Supervises Article 4 in the Member States?

National market surveillance authorities supervise Article 4 in the Member States, according to the Commission AI Literacy Q&A.

The national competent authorities differ by Member State, and their designation status is dated per country. Under Article 70, Member States shall designate at least one notifying authority and at least one market surveillance authority, designate a market surveillance authority as single point of contact, and make their contact information public by 2 August 2025.

Frequently Asked Questions about AI Literacy under Article 4

What does Article 4 require of providers and deployers?

Measures that support the development of AI literacy of their staff and of other persons dealing with AI systems on their behalf, calibrated to their knowledge, the context of use and the persons affected, without guaranteeing any individual's level.

What is the meaning of AI literacy?

Under Article 3(56), AI literacy means the skills, knowledge and understanding that allow providers, deployers and affected persons to make an informed deployment of AI systems and to gain awareness of the opportunities, risks and possible harm of AI.

Did the 2026 amendment abolish the AI literacy obligation?

No. Regulation (EU) 2026/1744 kept the duty on providers and deployers. The proposal COM(2025) 836 of 19 November 2025 would have turned it into a duty of the Commission and the Member States to encourage them, but that version was not adopted.

Is there a fine for breaching Article 4?

The AI Act sets no amount. Article 4 is not listed in Article 99(4), so any penalty depends on national rules under Article 99(1), which must be effective, proportionate and dissuasive.

Does Article 4 require a certificate?

No. The Commission AI Literacy Q&A states: "There is no need for a certificate. Organisations can keep an internal record of trainings and/or other guiding initiatives." Article 4(1) prescribes no format for that record.

Since when does the AI literacy obligation apply?

Since 2 February 2025. The amended wording, which asks for measures supporting AI literacy instead of a sufficient level, applies since 27 July 2026, when Regulation (EU) 2026/1744 entered into force.