New signups are temporarily closed.Existing customers can still sign in. Contact us to be notified when we reopen.
Skip to main content
Documentation

Governance & evidence

AI literacy

AI literacy is the obligation most organisations find out about late. It is short, it applies broadly, and unlike the high-risk regime it is already in force.

Transcript

Article 4 applies to nearly everyone deploying AI, it has been in force since February 2025, and it is the obligation most organisations discover last.

Open Settings and go to AI Literacy. This is the staff training centre for the Article 4 duty.

Start the assessment. You are asked a series of questions that test whether your people understand what AI systems are, how they are used, and the risks that come with them.

Answer each question, stepping through the assessment. A score of sixty percent or better earns the certificate.

Submit your answers. The platform marks the assessment against the Article 4 standard and records the result against you.

Pass, and you get a dated AI literacy certificate. It names Article 4 of the EU AI Act, so you have proof the training happened.

Retake it any time, and encourage your team to do the same. Each completed assessment is evidence that your organisation took the AI literacy duty seriously.

Why this is required

Article 4 requires providers and deployers of AI systems to take measures to support the development of AI literacy among their staff and other persons dealing with the operation and use of AI systems on their behalf. It does not require guaranteeing any specific level of literacy for any individual (as amended by the Digital Omnibus on AI).

Critically, it is not limited to high-risk systems. If your organisation deploys any AI system, Article 4 applies, and it has applied since 2 February 2025, alongside the Article 5 prohibitions.

The standard is contextual rather than fixed. Article 4 requires you to take into account your people's technical knowledge, experience, education and training, the context the systems are to be used in, and the persons or groups the systems are to be used on. Literacy for an engineering team differs from literacy for a recruitment team using a screening tool.

Article 3(56) defines AI literacy as 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 and risks and the possible harm it can cause.

There is a practical reason beyond compliance. Article 14 human oversight only works if the person overseeing the system understands its limits well enough to disagree with it. An overseer who cannot recognise a wrong answer provides no oversight at all, and automation bias is explicitly called out in Article 14(4)(b).

What EuroCompliant does

The literacy module assesses your organisation's current understanding of AI systems, risks, obligations and governance.

It records who has been trained, on what, and when, which is the documentation Article 4 measures actually produce. Each team member completes their own assessment individually, self-served from their own My Compliance page, rather than one shared result standing in for the whole organisation.

Results feed the governance documentation and evidence pack.

Deadlines

2 February 2025Article 4 has been in force since this date. It is not waiting for the high-risk deadline (now 2 December 2027).

Walking through it

1

Identify who is in scope

Staff and anyone else operating AI systems on your behalf, including contractors. Not just technical teams: the people using AI outputs to make decisions are squarely in scope.

Open /settings/literacy →
2

Each person takes their own assessment

From My Compliance, or directly from the literacy page: either way, the result is recorded against that specific person.

Open /my-compliance →
3

Assess current understanding

Article 4 asks you to account for existing knowledge and training, so establish the baseline before designing the intervention.

4

Tailor training to role and risk

Match the depth to the context and to the people the systems are used on. A team making decisions about individuals needs more than general awareness.

5

Record the measures taken

Article 4 requires measures, not certificates. The deliverable is a record showing you identified what people needed to know and acted on it.

The law

Frequently asked

Does Article 4 apply if we only use ChatGPT?

Yes. Article 4 applies to deployers of any AI system, with no high-risk threshold. Using a general-purpose assistant in a professional capacity puts you in scope.

Is there a certification we need?

No. Article 4 requires measures appropriate to your context, not an accredited qualification. What you need is a defensible record of what you assessed, what training you provided, and to whom.

Related guides

Try it on your own systems

Everything in this guide runs in the live product. Start a free trial and follow along with your own data.

Start free trial