Blend
/
Enforcement begins August 2, 2026

Is Your Organisation Ready for Article 4?

The EU AI Act requires all organisations deploying AI to ensure "sufficient AI literacy" among their staff. 7 questions. 2 minutes. Find out where you stand.

Question 1 of 70% complete

How many of your employees use AI tools in their daily work?

What is Article 4?

Article 4 of the EU AI Act requires all organisations that deploy or provide AI systems to ensure their staff have "sufficient AI literacy." This applies to every company operating in the EU that uses AI tools — from ChatGPT to automated recruitment systems.

Enforcement begins August 2, 2026. National market surveillance authorities will assess whether organisations can demonstrate that their AI literacy training is proportionate to the context, role, and risk level of AI use.

The key word is "sufficient." Generic awareness training may not meet the standard. Regulators will look for evidence that training was role-appropriate, measured for comprehension, and proportionate to the risk posed by the AI systems in use.

Built by Blend Training

We design scenario-based compliance training where people practice real decisions under pressure — not click-through slides. Our AI Act course puts learners in the role of HR directors, marketing leads, and CTOs facing actual compliance dilemmas.

Try the free demo View the full course

Your assessment scored "Gap" or "Critical"?

Book a 20-minute call. We'll walk through your specific gaps, show you the course module closest to your risk profile, and quote you a pilot price. No sales pitch — most calls end with a clear 3-step action list whether you work with us or not.

Book 20-min call →

Frequently asked questions about Article 4

What is Article 4 of the EU AI Act?
Article 4 requires every provider and deployer of AI systems operating in the EU to ensure their staff, contractors, and any third party operating an AI system on their behalf have a 'sufficient level of AI literacy.' It is a horizontal duty that applies regardless of the risk tier of the AI system being used. Even if the system is low-risk, the people running it must understand it.
Who has to comply with Article 4?
Any organisation that provides or deploys an AI system that is placed on the EU market or whose output is used in the EU. That includes EU companies, EU subsidiaries of non-EU companies, and non-EU companies whose AI system output reaches users in the EU. Practically: if your staff use ChatGPT, Microsoft Copilot, automated CV screening, AI-driven customer support, or any other AI system in the EU, Article 4 applies to you.
When does Article 4 enforcement begin?
Article 4 has been applicable since 2 February 2025, but national market surveillance authorities and the AI Office formally begin enforcement on 2 August 2026. From that date, supervisory authorities can investigate, request evidence, and impose corrective measures. Penalties under Article 99 reach €15 million or 3% of global annual turnover, whichever is higher.
What counts as 'sufficient' AI literacy training?
The Act does not prescribe a fixed syllabus. 'Sufficient' is judged against the context, role, and risk level of the AI system in use. A marketing team using ChatGPT for copy needs different training from a HR team using an AI shortlisting tool, which needs different training again from a clinical team using a diagnostic AI. The common floor that regulators are expected to look for: capability and limitations of the AI system, risks specific to the use case, the right of affected persons to explanation, when to escalate, and the organisation's internal policies and oversight mechanisms.
Does generic AI awareness training meet Article 4?
Generic awareness training is unlikely to satisfy a supervisory authority on its own. The text of Article 4 specifically references training that is 'taking into account their technical knowledge, experience, education and training, the context the AI systems are to be used in, and the persons or groups of persons on whom the AI systems are to be used.' That language anticipates role-tailored, context-specific training rather than a one-size-fits-all module.
What evidence will market surveillance authorities look for?
At a minimum: a documented training plan tied to the organisation's inventory of AI systems, evidence of which staff completed which training and when, evidence that training was proportionate to the role and risk, and an internal AI use policy. Best-practice evidence adds per-learner score or decision-path records and a refresh cycle aligned to material changes in the AI systems in use.
How often do staff need to be retrained?
The Act does not specify a refresh interval. Regulators are expected to assess whether training remains current relative to the AI systems in use. A pragmatic floor is annual refresh, with mandatory retraining when a material change occurs: new AI system introduced, significant model update, expansion of use case, or a regulatory or guidance change.
How does Article 4 interact with the rest of the EU AI Act?
Article 4 is the AI literacy obligation. Other articles impose separate duties: Article 5 prohibits specified AI practices (in force from 2 February 2025), Articles 6-49 govern high-risk AI systems (most provisions in force from 2 August 2026), Articles 50-56 cover transparency obligations and general-purpose AI models, and Articles 99-101 set the penalty framework. Article 4 training should reference the parts of the Act that are relevant to the AI systems the trainee uses.