Training your teams on AI in Monaco: AI Act Article 4 and a working method

Does a Monaco company have to train its teams on AI?
A Monaco company that falls within the scope of the AI Act, which can happen when it works with the European Union, must take AI literacy measures; formal training is not required in every case.1 Article 4 asks providers and deployers of AI systems to take measures to support the AI literacy of their staff, without demanding a set level.2 A Monaco company is covered if it places a system on the market or puts it into service in the Union, or if its system's output is used there.3 In every case, Law No. 1.565 governs the personal data a team hands to these tools.4
Key points
- The AI literacy duty has applied since 2 February 2025. Regulation (EU) 2026/1744 softened it: companies must support AI literacy, without guaranteeing a given level for each person.2
- Monaco is outside the Union, but the AI Act can reach a Monaco company that places a system on the market or puts it into service in the Union, or whose system produces output used in the Union.5,3
- In Monaco, FlashLearn AI is free for businesses based in the Principality, and Extended Monaco also runs the Digital FlashUP sessions and Digital LabImpact workshops.6,7
- Good training is judged by what people can do on their real tasks: check an answer, protect a piece of data, know when to take back control.
This guide is for the executive or team lead who wants to know what the law really asks, what Monaco makes available and how to organise training that is useful. It does not replace a lawyer's advice on your situation.
What AI Act Article 4 says since July 2026
The EU AI Regulation defines AI literacy as the skills, knowledge and understanding that allow people to deploy these systems in an informed way and to be aware of their opportunities, their risks and the harm they can cause.3 Article 4 belongs to Chapter I, which has applied since 2 February 2025.3,2
Its original version asked companies to ensure, to the best extent possible, a sufficient level of AI literacy.3 Regulation (EU) 2026/1744 of 8 July 2026, published in the Official Journal on 24 July and in force on the third day after that, replaced this wording.2 Providers and deployers must now "take measures to support the development of AI literacy" of their staff and of other persons dealing with the operation and use of these systems on their behalf.2
The text adds that this obligation "does not require providers or deployers to guarantee any specific level of AI literacy of any individual".2 The duty therefore remains: the company has to act, taking into account each person's knowledge, the context of use and the people on whom the system is used.2 The Commission and the Member States must for their part support these efforts, particularly those of SMEs, and the Commission will publish practical examples of how to comply.2
When a Monaco company is concerned
Monaco is not part of the European Union: its government describes it as the only non-EU country integrated into the European customs territory, and the association agreement talks opened in 2015 were suspended in 2023.5 The AI Act therefore does not apply there as such. It can still reach a Monaco company through its extraterritorial scope, in two main cases.3
- It is a provider: it 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 in the Union.3
- It is a provider or deployer, and the system's output is used in the Union: a score, a recommendation or a text produced from Monaco serves, for example, a client or a team in France or Italy.3
A deployer is any person or organisation using an AI system under its own authority, except for personal non-professional activity.3 A company that equips its staff with an off-the-shelf assistant can therefore be one without having developed anything. The Commission confirms that using ChatGPT to write advertising or translate text falls under Article 4, and that employees should then be informed of the tool's specific risks, such as hallucinations.1 It also reads the territorial scope broadly: the framework applies when the system is placed on the Union market, used in the Union, or when its use has an impact on people located there.1
A company serving only Monaco clients, with no output used in the Union and no impact on people located there, may stay outside the scope. This is checked activity by activity: our guide to the AI Act for Monaco companies sets out that analysis.
What the Commission expects in practice
The Commission has published a set of questions and answers on Article 4, updated on 20 October 2026. It sets no standard programme and lists four steps to consider at a minimum.1
- Ensure a general understanding: what is AI, how does it work, which tools do we use, what are their opportunities and dangers?1
- Establish your role: does the company develop systems or use someone else's?1
- Assess the risk of the systems: what do the employees using them need to know, and how can those risks be reduced?1
- Build the actions on this analysis, according to each person's level, the sector and the tool's purpose.1
Three clarifications avoid needless spending. No certificate is required: the company can keep an internal record of its training and guidance. No particular governance structure is imposed to comply with Article 4. Different levels of training for different people can be appropriate.1
Two limits matter just as much. Simply having staff read the instructions for use may, in many cases, be ineffective. Service providers using these systems on your behalf also need skills suited to their task.1 For high-risk systems, Article 26 goes further: human oversight must be assigned to people with the necessary competence, training and authority, and the necessary support.3 For those in Annex III, these rules apply from 2 December 2027.2
Supervision of Article 4 has been in the hands of national market surveillance authorities since 2 August 2026. The Commission stresses that penalties must be proportionate, and says a penalty might be more likely where an incident due to a lack of training and guidance is proven.1
What Monaco makes available to teams
The Prince's Government's Extended Monaco Businesses programme brings together several useful schemes, to be combined according to the team's level.8
- FlashLearn AI ("AI FlashLearn" on the programme's general page) is an online course of short videos in three chapters: "What is AI?", "Boosting productivity with AI" and "Ethics and AI". It is free for businesses based in Monaco. The access request includes the company's RCI number, and the applicant receives personal login details once it is approved.6,9 It is aimed at leaders, managers, employees and self-employed people.8
- Digital FlashUP offers awareness sessions of about 45 minutes.7 Its "AI & Business" series notably covered first uses on 4 March 2026, then, on 30 April 2026, bias, hallucinations, loss of control and the AI Act.10,9,11
- The AI questionnaire, named "AI Feasibility Questionnaire" on the AI programme page and "AI OpportunityA" on the general page, helps identify where AI can add value to the business.12,8
- Digital LabImpact brings small groups together in practical workshops to identify, test and integrate generative AI solutions.13
These schemes give a common foundation. Their pages do not present them as proof of compliance with Article 4: that foundation still has to be connected to your tools, your data and your trades. Our article on Extended Monaco and AI covers the whole programme.
Train by role, starting from real tasks
In France, among companies with 10 or more employees in mainly market sectors (excluding finance and agriculture) that use AI, 53% say a lack of expertise holds them back.14 In the projects we see, one identical course for everyone answers that gap poorly: an executive, a salesperson and an accountant do not take the same risks. Start by listing who uses which tool for which task, then prepare short content for each role.
- Management: decide which uses are allowed, who approves what and which tools are cleared, and keep the training record.
- Client-facing teams: prepare a reply with AI, spot invented information, never send a text nobody has reread. In a hotel, this applies to replies to guest requests.
- Sensitive functions (finance, compliance, human resources): know which decisions stay human and how to document the check. A bank or wealth manager adds its own secrecy and compliance rules.
The overall approach is set out in our guide on where to start with an AI transition. Our page on AI transition in Monaco explains how we support this shift within a team.
The data rules everyone should know
Pasting a client file containing personal data into an assistant is data processing. For a company established in Monaco, Law No. 1.565 applies to that gesture whether or not the AI Act is involved.4 Three reflexes follow for the whole team.
- Give the tool only what is needed: data must be adequate, relevant and limited to the purpose (Article 4 of the Law).4
- Use only the tools the company has approved: a processor must offer sufficient guarantees and process data only on documented instructions (Article 26).4
- Keep a person in the decision: everyone has the right not to be subject to a decision with legal or similarly significant effects based solely on automated processing, including profiling, subject to the exceptions in Article 19. In most of those exceptions, the person can obtain human intervention, express their point of view and contest the decision.4
To go further, read our guide on AI and personal data in Monaco.
Checking: the habit that matters most
A study presented at the CHI 2025 conference surveyed 319 knowledge workers on 936 real examples. The more people trust AI, the less critical thinking they report; the more confident they are in their own skills on the task, the more they apply it. Critical work shifts towards verifying information, integrating the answer and stewarding the task.15 These results are self-reported, but they point in a clear direction: train people in the job as much as in the tool.
A field experiment published in PNAS in 2025, with nearly a thousand high school students, points the same way. With open access to GPT-4, their grades rose by 48% during practice. Once the tool was removed, they were 17% lower than those of students who never had it. A tutor designed with safeguards largely limited this effect.16 A school setting does not transfer directly to a company, but the lesson is useful: a tool that does the work for someone can improve today's result without building the skill. Our article is AI making us stupid? explores this further.
Measure what people can do
An attendance rate tells you who came to a session, not what changed. We recommend checking a few observable actions, on a real task from the job, before the training and again a few weeks later.
- The person spots an error planted in an AI answer and traces it back to the source.
- They remove from a document the data that should not leave the company before using it.
- They can redo the core of the task without the tool and explain the result they approve.
- They know whom to tell about a doubt or an incident.
Note the date, audience, content and these checks in your internal record1: it is simple to keep, and it shows what the company has really put in place. Repeat the exercise whenever a tool or a use changes.
Frequently asked questions
Do we need a certificate to prove the team is trained on AI?
No. The Commission states that no certificate is needed for Article 4 and that an organisation can keep an internal record of its training and other guidance initiatives.1
Is FlashLearn AI enough to comply with Article 4?
It gives businesses based in Monaco a free foundation.6 Article 4 still requires taking account of people's level and the context of use,2 so your own tools, data and internal rules need to be added.
Is a company that only uses ChatGPT concerned?
If it falls within the scope of the AI Act, yes. The Commission gives the example of employees using ChatGPT to write advertising or translate text: they should be informed of the tool's specific risks, such as hallucinations.1
To find out which tasks and roles to train first, start from your real situation: our free AI audit reviews your uses, your tools and your data.
Request the free audit and prototype
Sources
- Commission européenne, AI Literacy: Questions & Answers,
- EUR-Lex, Journal officiel de l’Union européenne, Règlement (UE) 2026/1744 du 8 juillet 2026 (omnibus numérique sur l’IA),
- EUR-Lex, Journal officiel de l’Union européenne, Règlement (UE) 2024/1689 du 13 juin 2024 sur l’intelligence artificielle (AI Act),
- Légimonaco, Loi n° 1.565 du 3 décembre 2024 relative à la protection des données personnelles,
- Gouvernement Princier, Working with the European Union,
- Gouvernement Princier, Mon Entreprise, FlashLearn IA,
- Gouvernement Princier, Mon Entreprise, Develop your skills and take part in our events,
- Gouvernement Princier, Mon Entreprise, Extended Monaco Businesses program general information,
- Gouvernement Princier, Mon Entreprise, Digital FlashUP, Episode 3: AI & Business, AI, Ethics & Responsibilities,
- Gouvernement Princier, Mon Entreprise, Digital FlashUP, Episode 2: AI: first steps, initial uses,
- Gouvernement Princier, Mon Entreprise, Event replays,
- Gouvernement Princier, Mon Entreprise, AI programme,
- Gouvernement Princier, DIGITAL LabImpact Builds on the Momentum of DIGITAL FlashUP with Practical Workshops for Businesses,
- Insee, Les technologies de l’information et de la communication dans les entreprises en 2025, Insee Première n° 2120,
- ACM CHI 2025, Carnegie Mellon et Microsoft Research, Lee et al., The Impact of Generative AI on Critical Thinking,
- PNAS, Bastani et al., Generative AI without guardrails can harm learning: Evidence from high school mathematics,