At a glance
- Article 4 of the EU AI Act has applied since February 2, 2025. As amended by the Digital Omnibus on AI, Regulation (EU) 2026/1744, since July 27, 2026, it requires providers and deployers of AI systems to 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, without requiring them to guarantee any specific level of AI literacy of any individual.
- The Omnibus did not shift the duty to governments, as the Commission had proposed. It kept the duty on providers and deployers and added a duty on the Commission and Member States to support their efforts.
- Article 99 sets no fine tier for Article 4. Member States set the penalties and other enforcement measures for infringements of the Act, and the Commission says national market surveillance authorities enforce Article 4 from August 2, 2026, under national law.
- As of November 15, 2026, we found no US federal statute, and no state statute among the state laws this site covers, that requires private employers in general to train staff in AI literacy. Texas requires annual AI training for certain state and local government employees.
Is AI literacy legally required?
In the EU, yes, for any provider or deployer of an AI system within the AI Act’s reach: Article 4 requires 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. Since the Digital Omnibus, the obligation does not require them to guarantee any specific level of AI literacy of any individual. In the US, as of November 15, 2026, we found no federal statute, and no state statute among the state laws this site covers, that requires it of private employers in general. Specific rules require training for named roles, such as Texas public employees.
A company within the Act’s reach whose staff use a general-purpose chatbot for work is a deployer of that system and carries Article 4, even if nothing it does is high-risk.
What did Article 4 say before the Digital Omnibus?
As adopted in 2024, Article 4 read in full: “Providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level 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.”
It sits in Chapter I, which applies from February 2, 2025 (Article 113(a)). Article 3(56), which the Omnibus did not change, defines AI literacy as the “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.”
What did the Digital Omnibus change in Article 4?
Article 1(5) of Regulation (EU) 2026/1744 replaced Article 4 with three paragraphs. The text in force reads:
- “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.”
- “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).”
- “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.”
In paragraph 1, “ensure, to their best extent, a sufficient level of” became “support the development of,” and a new sentence says the obligation does not require them to guarantee any specific level for any individual. The rest is unchanged. Paragraphs 2 and 3 are new. Recital 8 gives the reason: stakeholder experience showed that stringent obligations to ensure a sufficient level would not suit all types of providers and deployers, and data indicated they create an additional compliance burden, particularly for smaller enterprises.
The Commission’s proposal, COM(2025) 836 of November 19, 2025, would have replaced Article 4 with: “The Commission and Member States shall encourage providers and deployers of AI systems to take measures to ensure a sufficient level of AI literacy” of their staff and others acting on their behalf. That would have ended the duty on companies. The adopted text kept it and lowered its standard.
The Omnibus sets no separate date for Article 4. It was published in the Official Journal on July 24, 2026 and entered into force on the third day after publication, July 27, 2026 (its Article 4). The original text therefore applied until July 27, 2026, and the amended text since. The Commission’s AI literacy pages say the Omnibus entered into force “in mid-July 2026”; the Official Journal gives July 27.
Who does Article 4 reach?
Providers and deployers of AI systems. A provider develops an AI system, or has one developed, and places it on the market or puts it into service under its own name or trademark (Article 3(3)). A deployer uses an AI system under its authority, except in the course of a personal non-professional activity (Article 3(4)).
The Act’s scope applies: providers placing AI systems on the EU market or putting them into service there, wherever established; deployers established or located in the EU; and providers and deployers in third countries where the system’s output is used in the EU (Article 2(1)(a) to (c)). Exclusions include, among others, AI systems used exclusively for military, defence or national security purposes, AI developed and put into service for the sole purpose of scientific research and development, and free and open-source AI systems unless placed on the market or put into service as high-risk or as a system under Article 5 or 50 (Article 2(3), (6) and (12)).
The Commission’s Q&A says the “other persons” are not employees but persons broadly under the organisational remit, such as a contractor, a service provider, or a client.
How is Article 4 enforced, and is there a fine?
There is no Article 4 fine tier. Article 99(3) covers the Article 5 prohibitions, 99(4) lists the operator obligations and notified body requirements in the 3% tier, and 99(5) covers incorrect, incomplete or misleading information supplied in reply to a request. Article 4 is on none of those lists.
That does not mean there is no sanction. Under amended Article 99(1), Member States lay down the rules on penalties and other enforcement measures, which may also include administrative fines, warnings and non-monetary measures, “applicable to any infringement of this Regulation by operators,” taking into account the interests of SMEs, including start-ups, and small mid-caps. The Commission’s Q&A says national market surveillance authorities, not the AI Office, supervise and enforce Article 4, from August 2, 2026, under national law. A sanction might be more likely, it says, if there is proof of an incident due to a lack of appropriate training and guidance. The AI Act does not create criminal offences or a right to compensation, though where there is harm a provider can be sued under national law.
What does the Commission’s guidance on AI literacy say?
The Commission’s AI Literacy Questions and Answers, last updated October 20, 2026, reflects the amended text and is guidance, not law. Its minimum steps for a program are items 2 to 4 of the program list. Its other points:
- No strict requirements or mandatory trainings are imposed. The requirements for training depend on the concrete context, and in many cases relying only on the instructions for use, or asking staff to read them, might be ineffective.
- Article 4 does not require measuring employees’ AI knowledge. No certificate is needed; organisations can keep an internal record of trainings or other guiding initiatives.
- No specific governance structure, such as an AI officer or governance board, is mandated to comply with Article 4.
- Staff who use a general chatbot to write advertising text or translate should be informed about specific risks, for example hallucination.
The Commission’s living repository of AI literacy practices held examples of more than 40 initiatives as of its July 27, 2026 page update. For now, replicating them does not automatically grant a presumption of compliance with Article 4, and the Commission says it will use the repository for the practical examples Article 4(2) requires. The Commission’s AI talent page, last updated October 13, 2026, says it is working with Member States in the AI Board on the Article 4(3) recommendations; we found none adopted as of November 15, 2026.
Which AI Act training duties sit outside Article 4?
Two high-risk provisions set competence duties the Omnibus left unchanged, as the Commission’s Q&A confirms for deployers.
- Article 26(2): “Deployers shall assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support.”
- Article 14(4): the provider must supply a high-risk system so that the people assigned to oversee it are enabled, as appropriate and proportionate, to understand its relevant capacities and limitations and monitor it, stay aware of automation bias, interpret its output, decide not to use it or disregard, override or reverse its output, and interrupt it. Article 14(5) adds two-person verification for Annex III point 1(a) remote biometric identification, with an exception.
Both apply from December 2, 2027 for stand-alone Annex III systems and August 2, 2028 for AI in Annex I Section A regulated products (Article 113(c), as amended).
Does US law require AI literacy training?
Not for private employers in general. As of November 15, 2026, we found no federal statute, and no state statute among the state laws this site covers, that does. The table below gives each narrower rule’s scope.
- Federal. Executive Order 14277 of April 23, 2025 states a policy “to promote AI literacy and proficiency among Americans” and directs agencies on K-12 resources, educator training, and AI apprenticeships. It creates no right or benefit enforceable at law or in equity (Section 9(c)) and contains no training duty for private employers.
- Federal agencies. For high-impact AI, OMB Memorandum M-25-21 requires agencies to ensure “sufficient and periodic training, assessment, and oversight for operators of AI to interpret and act on the AI’s output and manage associated risks” (Section 4(b)(iv)).
- Texas. HB 3512, effective September 1, 2025, requires identified state agency employees and officers, and identified local government employees and officials, to complete a certified AI training program at least once a year (Government Code 2054.5191). HB 149 adds no private-sector training duty.
- Illinois. Public Act 104-0399 requires the State Board of Education to publish K-12 AI guidance, including best practices for developing student literacy in AI, by July 1, 2026.
- Colorado. SB 26-189 has no general training duty, but a reviewer providing meaningful human review must be “trained to conduct the review” (C.R.S. 6-1-1701(15)). A federal court order in X.AI LLC v. Weiser limits Attorney General enforcement; the Colorado guide gives its terms and status.
Some guidance asks for role-based AI training without calling it literacy. NYDFS Insurance Circular Letter No. 7, guidance for insurers authorized in New York that use AI systems or external consumer data in underwriting or pricing, says their policies and procedures should include training for relevant personnel. NIST’s voluntary AI RMF (GOVERN 2.2) calls for AI risk management training. The AI acceptable use policy guide quotes both.
| Rule | Who it reaches | What training or literacy it requires | Binding or voluntary | Date |
|---|---|---|---|---|
| EU AI Act Article 4(1), as amended by Regulation (EU) 2026/1744 | Providers and deployers of AI systems within Article 2 | Measures to support the development of AI literacy of 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, the context of use, and the persons or groups on whom the systems are used; the obligation does not require providers or deployers to guarantee any specific level of AI literacy of any individual | Binding; no Article 99 fine tier; penalties and other enforcement measures under national rules (Article 99(1)) | Applied since February 2, 2025; amended text since July 27, 2026 |
| EU AI Act Article 4(2) and (3) | The Commission, Member States, and the AI Board | Support and facilitate providers’ and deployers’ efforts, in particular SMEs; Commission publishes practical examples on the single information platform; Board adopts recommendations with common objectives | Binding on the Commission, Member States and Board | Since July 27, 2026; no Board recommendations found as of November 15, 2026 |
| EU AI Act Article 26(2) | Deployers of high-risk AI systems | Human oversight assigned to natural persons who have the necessary competence, training and authority, as well as the necessary support | Binding; Article 99(4)(e) tier | December 2, 2027 (Annex III); August 2, 2028 (Annex I Section A) |
| EU AI Act Article 14(4) and (5) | Providers of high-risk AI systems (design); deployers of Annex III point 1(a) systems (verification) | Systems provided so overseers are enabled, as appropriate and proportionate, to understand, monitor, interpret, override and stop them; for Annex III point 1(a), no action or decision on an identification unless separately verified and confirmed by at least two natural persons with the necessary competence, training and authority, except for law enforcement, migration, border control or asylum uses where Union or national law considers this disproportionate | Binding | Same dates as Article 26 |
| Commission AI Literacy Q&A and living repository | Providers and deployers | Minimum steps for a program; internal records; for now, repository practices carry no automatic presumption of compliance | Guidance | Q&A last updated October 20, 2026 |
| Executive Order 14277 (90 FR 17519) | Federal agencies, chiefly those on the White House Task Force on Artificial Intelligence Education, and all agencies that provide educational grants (Section 8(e)) | K-12 AI literacy resources, educator training priorities in grant programs, AI-related Registered Apprenticeships | Directs agencies; creates no enforceable right or benefit | Signed April 23, 2025 |
| OMB Memorandum M-25-21 | Federal agencies | The federal workforce has a responsibility to develop and maintain, at a minimum, foundational knowledge of how to use AI responsibly (Section 2(e)); for high-impact AI, sufficient and periodic training, assessment, and oversight for operators, specific to the system and its use (Section 4(b)(iv)) | Binding on agencies | Issued April 3, 2025 |
| AI Training Act (Public Law 117-207, 41 U.S.C. 1703 note) | The Director of OMB, for the covered workforce: executive agency employees responsible for program management; the planning, research, development, engineering, testing, and evaluation of systems; procurement and contracting; logistics; or cost estimating, and other personnel the agency head designates (the Department of Defense and the NNSA are excluded) | An AI training program, provided within one year of enactment and at least annually after that, so that the covered workforce has knowledge of the capabilities and risks associated with AI; topics include how AI works, its benefits to the Federal Government, its risks, including discrimination and risks to privacy, ways to mitigate them, and future trends; updated at least every two years | Binding on OMB; no duty on private employers | Enacted October 17, 2022; the section has no force or effect from October 17, 2032 |
| Texas HB 3512 (Government Code 2054.5191, 2054.5193) | State agency employees who use a computer for at least 25 percent of their duties and are identified by the agency, and agency officers; local government employees and officials with access to a local government computer system or database who use a computer for at least 25 percent of their duties | A certified AI training program at least once a year; certified programs must focus on how AI may be used in relation to a state employee’s duties and teach best practices on literacy in deploying and operating AI; in a school district, only the district’s cybersecurity coordinator must complete the training annually, and other employees required to complete it do so as the district determines (Education Code 11.175(h-1)); employees on certain leave are excluded (2054.5191(f)) | Binding | Effective September 1, 2025 |
| Texas HB 149 (Business and Commerce Code Sec. 554.102, added by the Act) | The Texas Artificial Intelligence Council | Conduct training programs for state agencies and local governments on the use of AI systems; no training duty on private persons | Binding on the council | Effective January 1, 2026 |
| Illinois Public Act 104-0399 (105 ILCS 5/2-3.118a) | Illinois State Board of Education | Statewide guidance for school districts and educators on AI in elementary and secondary education, including best practices for developing student literacy in AI | Binding on the State Board | Effective January 1, 2026; guidance due by July 1, 2026 |
| California AB 2876 (Statutes of 2024, Chapter 927; Education Code 33548) | California Instructional Quality Commission | Consider incorporating AI literacy content into the mathematics, science, and history-social science curriculum frameworks when they are next revised after January 1, 2025, and consider including AI literacy in its criteria for evaluating instructional materials when the State Board next adopts materials in those subjects; AI literacy is defined as the knowledge, skills, and attitudes associated with how AI works, including its principles, concepts, and applications, and how to use AI, including its limitations, implications, and ethical considerations | Binding on the commission to consider; no training duty | Effective January 1, 2025 |
| Colorado SB 26-189 (C.R.S. 6-1-1701(15), 6-1-1705(1)(a)(II)) | Deployers of covered ADMT | After an adverse outcome from a consequential decision that a covered ADMT materially influences, on the consumer’s request and to the extent commercially reasonable, an opportunity for meaningful human review by a designated individual with authority to approve, modify, or override the decision who is trained to conduct the review | Binding; Attorney General enforcement limited by the X.AI LLC v. Weiser order | Consequential decisions on or after January 1, 2027 |
What should an AI literacy program cover?
This list is a starting point, not a complete program for any one law or sector:
- Scope: every AI system provided or deployed, and every group that operates or uses one on the organisation’s behalf, including contractors and service providers (Article 4(1); Commission Q&A).
- Basics: what AI is, how it works, which systems the organisation uses, and their opportunities and dangers (Commission Q&A).
- Role and risk: provider or deployer for each system, and what staff need to know about its risks, such as hallucination (Article 3(3) and (4); Commission Q&A).
- Levels and context: depth set by technical knowledge, experience, education and training, and by the context of use and the people affected (Article 4(1)).
- Human oversight: named overseers of high-risk systems trained on capacities, limitations, automation bias, interpretation, override and stop (Articles 14(4) and 26(2)).
- Legal and ethical aspects of the systems in use (Commission Q&A).
- US role-specific training: Texas public employees (2054.5191), Colorado reviewers (6-1-1701(15)), federal high-impact AI operators (M-25-21).
- Records: a dated internal record of trainings and other initiatives, kept by group.
What should a company do now?
- List every AI system you provide or deploy, including chatbots staff use for work, and who uses each on your behalf
- Confirm which systems fall within Article 2, including third-country providers and deployers whose system output is used in the EU
- Take measures to support the development of AI literacy of staff and other persons dealing with the operation and use of AI systems on your behalf, taking into account their technical knowledge, experience, education and training, the context of use, and the people the systems are used on; Article 4 does not require you to guarantee any specific level for any individual
- Keep a dated internal record of trainings and other initiatives
- Assign human oversight of any high-risk system you deploy to people with the necessary competence, training and authority, and give them the necessary support, before December 2, 2027 (Annex III) or August 2, 2028 (Annex I Section A)
- Watch for the Article 4(2) practical examples and any Article 4(3) Board recommendations
- Check the national penalty rules in each Member State where you operate
How this fits the other guides: the EU AI Act compliance guide and the EU AI Act compliance checklist cover scope, the Omnibus dates and the other provider and deployer duties; the AI acceptable use policy guide covers the staff policy and its training section; the AI governance policy guide covers the organisation-level policy; and the Illinois AI laws guide covers the school and teacher-evaluation Acts.
This guide is general information as of November 15, 2026, not legal advice. Confirm how these rules apply to your organisation with counsel or the competent regulator.
Common questions
Is AI literacy training mandatory under the EU AI Act?
Article 4 requires providers and deployers of AI systems within the Act's reach to 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, and has applied since February 2, 2025. Since Regulation (EU) 2026/1744 amended it with effect from July 27, 2026, it does not require them to guarantee any specific level of AI literacy of any individual. The Commission's Q&A says no strict requirements or mandatory trainings are imposed, that the requirements for training depend on the concrete context, and that in many cases relying only on the instructions for use, or asking staff to read them, might be ineffective.
Did the Digital Omnibus remove the AI literacy obligation for companies?
No. The Commission's November 19, 2025 proposal, COM(2025) 836, would have replaced the duty with one on the Commission and Member States to encourage providers and deployers. The adopted Regulation (EU) 2026/1744 kept the duty on providers and deployers, changed it from taking measures to ensure, to their best extent, a sufficient level of AI literacy to taking measures to support its development, added that the obligation does not require them to guarantee any specific level of AI literacy of any individual, and required the Commission and Member States to support those efforts and the AI Board to adopt recommendations.
What is the fine for breaching Article 4 of the EU AI Act?
Article 99 sets no fine tier for Article 4: it is not listed in Article 99(3), (4) or (5). Under amended Article 99(1), Member States lay down the rules on penalties and other enforcement measures, which may also include administrative fines, warnings and non-monetary measures, applicable to any infringement of the Regulation by operators. The Commission says national market surveillance authorities, not the AI Office, supervise and enforce Article 4, from August 2, 2026, under national law.
By industry: all 8 industry pages
Sources
- Regulation (EU) 2024/1689 (Artificial Intelligence Act), as adopted, Articles 3(56), 4 and 113 and recital 20, Official Journal of the European Union, L series, July 12, 2024.
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), Article 1(5), (38) and (40), Article 4 and recital 8, Official Journal of the European Union, L series, July 24, 2026.
- Regulation (EU) 2024/1689, consolidated text as of July 27, 2026, Articles 2, 3, 4, 14, 26, 62, 99 and 113, EUR-Lex, accessed November 15, 2026.
- Proposal for a Regulation (Digital Omnibus on AI), COM(2025) 836 final, Article 1(4) and recital 5, European Commission, November 19, 2025.
- AI Literacy: Questions & Answers, European Commission, last updated October 20, 2026.
- AI talent, skills and literacy, European Commission, last updated October 13, 2026.
- Repository of AI literacy practices, European Commission, last updated July 27, 2026.
- Living repository to foster learning and exchange on AI literacy, European Commission, last updated August 10, 2026.
- Executive Order 14277, Advancing Artificial Intelligence Education for American Youth, 90 FR 17519, Federal Register, April 28, 2025.
- OMB Memorandum M-25-21, Accelerating Federal Use of AI through Innovation, Governance, and Public Trust, Office of Management and Budget, April 3, 2025.
- Public Law 117-207, Artificial Intelligence Training for the Acquisition Workforce Act (AI Training Act), 136 Stat. 2238, U.S. Government Publishing Office, approved October 17, 2022.
- H.B. No. 3512 (enrolled), relating to artificial intelligence training programs for certain employees and officials of state agencies and local governments, Texas Legislature, 2025.
- History for 89(R) HB 3512, Texas Legislature Online, accessed November 15, 2026.
- H.B. No. 149 (enrolled), Texas Responsible Artificial Intelligence Governance Act, Texas Legislature, 2025.
- Public Act 104-0399 (SB 1920), Illinois General Assembly, approved August 15, 2025, effective January 1, 2026.
- AB 2876 (Berman), Chapter 927, Statutes of 2024, amending Education Code Section 33548, California Legislature, approved September 29, 2024.
- SB 26-189, Session Laws 2026, Chapter 131 (enrolled act), Colorado General Assembly, approved May 14, 2026.
- Insurance Circular Letter No. 7 (2024): Use of Artificial Intelligence Systems and External Consumer Data and Information Sources in Insurance Underwriting and Pricing, New York State Department of Financial Services, July 11, 2024.
- Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1 (PDF), GOVERN 2.2, National Institute of Standards and Technology, January 26, 2023.