AI Literacy by Role: How Irish Employers Can Tailor AI Training (Article 4)

If you are looking at AI literacy by role in Ireland, the starting point is that EU law does not hand employers a single syllabus. Article 4 of the AI Act asks providers and deployers of AI systems to take measures on AI literacy, and the European Commission's own Q&A says the measures should reflect who the people are and how AI is used. That makes a role-based approach a natural fit for Dublin employers. This guide explains what the Commission's page says, and then offers some practical ways to tailor a programme. For the basic duty itself, see our explainer on the Article 4 AI literacy duty for employers.

What the sources say about tailoring

The Commission's "AI Literacy - Questions & Answers" page (last updated 27 July 2026) is the most useful official source on tailoring. According to that page, providers and deployers must take measures to support AI literacy for their staff and for "other persons dealing with the operation and use of AI systems on their behalf". The page says those measures should take account of the technical knowledge, experience, education and training of the people involved, and of the context in which the AI systems are used.

The Commission also states that no strict requirements or mandatory trainings are imposed and that there is no one-size-fits-all format. It says having different levels of training or learning approaches could be appropriate, and that organisations should consider their role and the risks of their systems, so high-risk systems may call for additional measures.

The consolidated EUR-Lex text of Regulation (EU) 2024/1689 (version dated 27 July 2026) now reads, in Article 4, that providers and deployers "shall take measures to support the development of AI literacy of their staff", and adds that this does not require them to guarantee any specific level of AI literacy of any individual. The Commission's page explains that the Digital Omnibus on AI amended Article 4, and that the amended wording no longer mandates a "sufficient" level. If you read older commentary, it may quote the earlier wording, so check which version it is using.

Why one generic course tends to miss the point

Because the law points to knowledge, experience, education and context, a single awareness session for everyone is hard to justify as the whole answer. The Commission's page gives examples that show why. Staff who use a general chatbot for advertising copy or translation should, it says, be told about specific risks such as hallucination. Staff who work with a human-in-the-loop system need skills specific to that system. And the page says that simply relying on a system's instructions for use, or asking staff to read them, "might be ineffective".

The page also says people with AI development degrees or experience are "normally" considered literate, but that organisations should still check that they know the organisation's own systems, risks and the legal and ethical aspects. In other words, even your technical staff may need a role-specific briefing.

A practical way to build role-based tiers

Our suggestion: start from what each group actually does with AI, not from job titles alone. The groupings below are our own illustration, not a legal classification, and the Commission's page does not prescribe any particular tiers.

• Everyone who may touch AI tools: a common baseline covering what the tools are, what they can get wrong, what company data must not be entered, and who to ask. The Commission's suggested minimum steps include building a general understanding of AI.
• Everyday users of generative tools: hands-on guidance for the tasks they perform, including checking outputs before relying on them and recognising hallucination.
• People who oversee or sign off AI-assisted decisions: deeper material on the specific system, its limits and when to override it. The Commission's page notes that for deployers of high-risk systems the duty to train staff for human oversight remains.
• Managers and policy owners: how the organisation decides which tools are allowed, who is accountable, and how incidents are escalated.
• Procurement, IT and data protection contacts: what to ask vendors, and how the organisation's role (provider or deployer) shapes its duties.

Work out your role first

The Commission's page recommends determining whether your organisation is a provider or a deployer, and assessing the risks of the AI systems you use, before tailoring actions to the audience and context. The Act defines a deployer as an organisation using an AI system under its authority (outside purely personal, non-professional activity). Our reading is that an employer buying in tools is likely to be a deployer, but check each tool and any in-house development.

Contractors and other people working on your behalf

The Commission's page says "other persons" covers people who are not employees but fall under the organisation's remit, such as contractors, service providers or clients, depending on the risk. It adds that contractors' staff need the appropriate AI skills for their tasks, and that whether a service provider must contractually demonstrate AI literacy depends on the type of system and the risk. A practical approach: list the third parties who operate or use AI on your behalf and decide, tool by tool, what you will ask of them.

HR teams as a distinct audience

HR is often where AI touches people's livelihoods directly, for instance in recruitment tools. The Commission describes employment as a high-risk area, and its AI Act page (last updated 3 August 2026) says those rules apply from 2 December 2027 following the Digital Omnibus. Dates have moved before, so check the current position. For a deeper look at this audience, see our guide to AI training for HR teams in Ireland.

Keeping a record

The Commission's page says no certificate is needed and that organisations can keep an internal record of training and other initiatives. It also says Article 4 does not require measuring employees' AI knowledge and that no AI officer or governance board is required. A practical approach: keep a short log noting the role group, the content covered, the date and who attended, so you can show what you did and why you chose it. That log is our suggestion, not a requirement stated on the page.

Enforcement and what to expect

According to the Commission, Article 4 entered into application on 2 February 2025, and supervision and enforcement lie with national market surveillance authorities, not the AI Office, with penalties depending on national laws. The page says penalties must be proportionate, and that sanctions might be more likely if there is proof of an incident due to lack of appropriate training. It gives slightly different start dates for supervision in different places (2 August 2026 and 3 August 2026), so check the page itself for the current position. The page does not name the Irish authority, so we have not done so here.

Paying for it

Role-based programmes cost time and money. Public supports exist and change over time. Our overview of funding supports for AI skills training explains where to look. Check the official pages for current eligibility before you plan around any scheme.

Next steps

• List the AI tools in use and who uses them.
• Decide your role (provider or deployer) for each.
• Group people by what they do with AI, not only by seniority.
• Match content to each group, including limits and data rules.
• Record what you did, and review when tools or the law change.

Learnsignal's offices are off Baggot Street in Dublin, and we offer in-person training from there. For current details, see our CPD pages.

Frequently asked questions

Does the AI Act require a certificate for AI literacy?

The Commission's Q&A page says no certificate is needed. It says organisations can keep an internal record of training and other initiatives.

Do all staff need the same AI training?

No. The Commission's page says there is no one-size-fits-all format and allows different levels for different audiences, depending on knowledge, experience and the context of use.

Does Article 4 cover contractors?

The Commission says the duty extends to other persons dealing with the operation and use of AI systems on the organisation's behalf, which can include contractors and service providers, depending on the risk.

Has Article 4 changed?

According to the Commission's page, the Digital Omnibus on AI amended it, so that the duty is to support the development of AI literacy and no longer mandates a "sufficient" level. Check the current consolidated text on EUR-Lex.

The right answer for your organisation depends on your tools, your people and your risks, so build the programme around those and review it regularly.